This paper proposes a fully data-driven approach for optimal control of nonlinear control-affine systems represented by a stochastic diffusion. The focus is on the scenario where both the nonlinear dynamics and stage cost functions are unknown, while only a control penalty function and constraints are provided. To this end, we embed state probability densities into a reproducing kernel Hilbert space (RKHS) to leverage recent advances in operator regression, thereby identifying Markov transition operators associated with controlled diffusion processes. This operator learning approach integrates naturally with convex operator-theoretic Hamilton-Jacobi-Bellman recursions that scale linearly with state dimensionality, effectively solving a wide range of nonlinear optimal control problems. Numerical results demonstrate its ability to address diverse nonlinear control tasks, including the depth regulation of an autonomous underwater vehicle.

In this paper, we address a practical distributed Bayesian learning problem with asynchronous measurements and predictions due to diverse computational conditions. To this end, asynchronous distributed Gaussian process (AsyncDGP) regression is proposed, which is the first effective online distributed Gaussian processes (GPs) approach to improve the prediction accuracy in real-time learning tasks. By leveraging the devised evaluation criterion and established prediction error bounds, AsyncDGP enables the distinction of contributions of each model for prediction ensembling using aggregation strategy. Furthermore, we extend its utility to dynamic systems by introducing a learning-based control law, ensuring guaranteed control performance in safety-critical applications. Additionally, a networked online learning simulation platform for distributed GPs, namely online GP gym (GPgym), is introduced for testing the performance of learning and control of dynamical systems. Numerical simulations within GPgym across regression tasks with real-world data sets and dynamical control scenarios demonstrate the effectiveness and applicability of AsyncDGP.

We propose a family of Gaussian processes (GP) for dynamical systems with linear time-invariant responses, which are nonlinear only in initial conditions. This linearity allows us to tractably quantify forecasting and representational uncertainty, simultaneously alleviating the challenge of computing the distribution of trajectories from a GP-based dynamical system and enabling a new probabilistic treatment of learning Koopman operator representations. Using a trajectory-based equivariance – which we refer to as Koopman equivariance – we obtain a GP model with enhanced generalization capabilities. To allow for large-scale regression, we equip our framework with variational inference based on suitable inducing points. Experiments demonstrate on-par and often better forecasting performance compared to kernel-based methods for learning dynamical systems.

The practical deployment of learning-based autonomous systems would greatly benefit from tools that flexibly obtain safety guarantees in the form of certificate functions from data. While the geometrical properties of such certificate functions are well understood, synthesizing them using machine learning techniques still remains a challenge. To mitigate this issue, we propose a diffeomorphic function learning framework where prior structural knowledge of the desired output is encoded in the geometry of a simple surrogate function, which is subsequently augmented through an expressive, topology-preserving state-space transformation. Thereby, we achieve an indirect function approximation framework that is guaranteed to remain in the desired hypothesis space. To this end, we introduce a novel approach to construct diffeomorphic maps based on RBF networks, which facilitate precise, local transformations around data. Finally, we demonstrate our approach by learning diffeomorphic Lyapunov functions from real-world data and apply our method to different attractor systems.

Control tasks with safety requirements under high levels of model uncertainty are increasingly common. Machine learning techniques are frequently used to address such tasks, typically by leveraging model error bounds to specify robust constraint-based safety filters. However, if the learned model uncertainty is very high, the corresponding filters are potentially invalid, meaning no control input satisfies the constraints imposed by the safety filter. While most works address this issue by assuming some form of safe backup controller, ours tackles it by collecting additional data on the fly using a Gaussian process bandit-type algorithm. We combine a control barrier function with a learned model to specify a robust certificate that ensures safety if feasible. Whenever infeasibility occurs, we leverage the control barrier function to guide exploration, ensuring the collected data contributes toward the closed-loop system safety. By combining a safety filter with exploration in this manner, our method provably achieves safety in a general setting that does not require any prior model or backup controller, provided that the true system lies in a reproducing kernel Hilbert space. To the best of our knowledge, it is the first safe learning-based control method that achieves this.

Handling orientations of robots and objects is a crucial aspect of many applications. Yet, ever so often, there is a lack of mathematical correctness when dealing with orientations, especially in learning pipelines involving, for example, artificial neural networks. In this paper, we investigate reinforcement learning with orientations and propose a simple modification of the network's input and output that adheres to the Lie group structure of orientations. As a result, we obtain an easy and efficient implementation that is directly usable with existing learning libraries and achieves significantly better performance than other common orientation representations. We briefly introduce Lie theory specifically for orientations in robotics to motivate and outline our approach. Subsequently, a thorough empirical evaluation of different combinations of orientation representations for states and actions demonstrates the superior performance of our proposed approach in different scenarios, including: direct orientation control, end effector orientation control, and pick-and-place tasks.

This paper proposes a novel distributed coverage controller for a multi-agent system with constant-speed unicycle robots (CSUR). The work is motivated by the limitation of the conventional method that does not ensure the satisfaction of hard state-and input-dependent constraints and leads to feasibility issues for multi-CSUR systems. In this paper, we solve these problems by designing a novel coverage cost function and a saturated gradient-search-based control law. Theoretical proofs are provided to guarantee that the CSURs ultimately move to the optimal coverage configuration without moving out of the covered domain. The controller is implemented in a distributed manner based on a novel communication standard among the agents. A series of simulation studies are conducted to validate the correctness of our theory by showing the efficacy of the proposed coverage controller in different initial conditions and with various control parameters. A comparison study in simulation reveals the advantage of the proposed method over the conventional method in terms of avoiding infeasibility. The experimental study verifies the applicability of the method to real robots. The development procedure of the method from theoretical analysis to experimental validation provides a novel framework for multi-agent system coordinate control with complex dynamics.Note to Practitioners—This paper gives a novel method to effectively cover a polygonal area using multiple constant-speed unicycle robots (CSUR) like wheeled robots and fixed-wing unmanned aerial vehicles (fUAV). Compared to the conventional approaches, our method allows these robots to cover a target region using circular orbits without departing the covered region. Also, the method satisfies common control saturation constraints in practice and can be implemented in a reliable distributed scheme. While the efficacy and correctness of the proposed method are rigorously proved using control theory, we also provide necessary interpretive elucidations to explain its underlying mechanism and selection rationale. The method is validated to be effective for wheeled robots in experimental studies, although it can also be applied to fUAVs in theory.

In the realm of the cooperative control of multiagent systems (MASs) with unknown dynamics, Gaussian process (GP) regression is widely used to infer the uncertainties due to its modeling flexibility of nonlinear functions and the existence of a theoretical prediction error bound. Online learning, which involves incorporating newly acquired training data into GP models, promises to improve control performance by enhancing predictions during the operation. Therefore, this article investigates the online cooperative learning algorithm for MAS control. Moreover, an event-triggered data selection mechanism, inspired by the analysis of a centralized event-trigger (CET), is introduced to reduce the model update frequency and enhance the data efficiency. With the proposed learning-based control, the practical convergence of the MAS is validated with guaranteed tracking performance via the Lyapunov theory. Furthermore, the exclusion of the Zeno behavior for individual agents is shown. Finally, the effectiveness of the proposed event-triggered online learning method is demonstrated in simulations.

A drawing robot avatar is a robotic system that allows for telepresence-based drawing, enabling users to remotely control a robotic arm and create drawings in real-time from a remote location. The proposed control framework aims to improve bimanual robot telepresence quality by reducing the user workload and required prior knowledge through the automation of secondary or auxiliary tasks. The introduced novel method calculates the near-optimal Cartesian end-effector pose in terms of visual feedback quality for the attached eye-to-hand camera with motion constraints in consideration. The effectiveness is demonstrated by conducting user studies of drawing reference shapes using the implemented robot avatar compared to stationary and teleoperated camera pose conditions. Our results demonstrate that the proposed control framework offers improved visual feedback quality and drawing performance.

Ensuring safety and adapting to the user’s behavior are of paramount importance in physical human-robot interaction. Thus, incorporating elastic actuators in the robot’s mechanical design has become popular, since it offers intrinsic compliance and additionally provide a coarse estimate for the interaction force by measuring the deformation of the elastic components. While observer-based methods have been shown to improve these estimates, they rely on accurate models of the system, which are challenging to obtain in complex operating environments. In this work, we overcome this issue by learning the unknown dynamics components using Gaussian process (GP) regression. By employing the learned model in a Bayesian filtering framework, we improve the estimation accuracy and additionally obtain an observer that explicitly considers local model uncertainty in the confidence measure of the state estimate. Furthermore, we derive guaranteed estimation error bounds, thus, facilitating the use in safety-critical applications. We demonstrate the effectiveness of the proposed approach experimentally in a human-exoskeleton interaction scenario.

Control over networks is envisioned to be one of the driving applications of future mobile networks. Networked control systems contain sensors and controllers exchanging time-sensitive information to fulfill a particular control goal. In this work, we consider$N$heterogeneous feedback control loops closed over a wireless star network. A centralized scheduler located at the central node, i.e., base station (BS), determines the transmission schedule of sensor-to-BS and BS-to-controller communication links. We assume that each link can accommodate a single transmission at a time and is prone to data losses with time-varying probability. Moreover, each controller estimates the system state remotely based on available information. In such a setting, we formulate an optimization problem to minimize the network-induced estimation error at the controller. In particular, we determine the optimal transmission schedule on each link that leads to the minimum normalized mean squared error (nMSE) in a given finite horizon (FH). We compare the performance of our proposed FH scheduler to various schedulers from the existing literature. Our simulation results show that by solving the finite horizon problem optimally, we are able to reduce the nMSE by$10\%$when compared to the best performing scheduling policy among the selected policies from the state-of-the-art. Moreover, the linear-quadratic Gaussian (LQG) cost is reduced by more than$13\%$indicating a control performance improvement in the network.

This article proposes a data-driven modeling framework with physically consistent Gaussian Processes (GPs), enabling learning-based disturbance estimation for uncertain mechanical systems with covariance-adaptive Momentum Observers (MOs). We present novel error bound results in closed form, holding with high, exactly computable probabilities, and exploit the multidimensional, physically constrained function distribution induced by the differential equation structure of Lagrangian systems. The inherent uncertainty quantification provided by GPs and the derived model error bounds are then leveraged to probabilistically guarantee exponential stability of a class of data-driven, adaptive MOs with user-definable convergence parameters. We demonstrate the performance of our proposed methods in simulations and physical experiments, showing significant improvements compared to the state-of-the-art from industry and research.

Many machine learning approaches for decision making, such as reinforcement learning, rely on simulators or predictive models to forecast the time-evolution of quantities of interest, e.g., the state of an agent or the reward of a policy. Forecasts of such complex phenomena are commonly described by highly nonlinear dynamical systems, making their use in optimization-based decision-making challenging. Koopman operator theory offers a beneficial paradigm for addressing this problem by characterizing forecasts via linear time-invariant (LTI) ODEs, turning multi-step forecasts into sparse matrix multiplication. Though there exists a variety of learning approaches, they usually lack crucial learning-theoretic guarantees, making the behavior of the obtained models with increasing data and dimensionality unclear. We address the aforementioned by deriving a universal Koopman-invariant reproducing kernel Hilbert space (RKHS) that solely spans transformations into LTI dynamical systems. The resulting Koopman Kernel Regression (KKR) framework enables the use of statistical learning tools from function approximation for novel convergence results and generalization error bounds under weaker assumptions than existing work. Our experiments demonstrate superior forecasting performance compared to Koopman operator and sequential data predictors in RKHS.

While Gaussian processes are a mainstay for various engineering and scientific applications, the uncertainty estimates don't satisfy frequentist guarantees and can be miscalibrated in practice. State-of-the-art approaches for designing calibrated models rely on inflating the Gaussian process posterior variance, which yields confidence intervals that are potentially too coarse. To remedy this, we present a calibration approach that generates predictive quantiles using a computation inspired by the vanilla Gaussian process posterior variance but using a different set of hyperparameters chosen to satisfy an empirical calibration constraint. This results in a calibration approach that is considerably more flexible than existing approaches, which we optimize to yield tight predictive quantiles. Our approach is shown to yield a calibrated model under reasonable assumptions. Furthermore, it outperforms existing approaches in sharpness when employed for calibrated regression.

Gaussian processes have become a promising tool for various safety-critical settings, since the posterior variance can be used to directly estimate the model error and quantify risk. However, state-of-the-art techniques for safety-critical settings hinge on the assumption that the kernel hyperparameters are known, which does not apply in general. To mitigate this, we introduce robust Gaussian process uniform error bounds in settings with unknown hyperparameters. Our approach computes a confidence region in the space of hyperparameters, which enables us to obtain a probabilistic upper bound for the model error of a Gaussian process with arbitrary hyperparameters. We do not require to know any bounds for the hyperparameters a priori, which is an assumption commonly found in related work. Instead, we are able to derive bounds from data in an intuitive fashion. We additionally employ the proposed technique to derive performance guarantees for a class of learning-based control problems. Experiments show that the bound performs significantly better than vanilla and fully Bayesian Gaussian processes.

The safe operation of physical systems typically relies on high-quality models. Since a continuous stream of data is generated during run-time, such models are often obtained through the application of Gaussian process regression because it provides guarantees on the prediction error. Due to its high computational complexity, Gaussian process regression must be used offline on batches of data, which prevents applications, where a fast adaptation through online learning is necessary to ensure safety. In order to overcome this issue, we propose the LoG-GP. It achieves a logarithmic update and prediction complexity in the number of training points through the aggregation of locally active Gaussian process models. Under weak assumptions on the aggregation scheme, it inherits safety guarantees from exact Gaussian process regression. These theoretical advantages are exemplarily exploited in the design of a safe and data-efficient, online-learning control policy. The efficiency and performance of the proposed real-time learning approach is demonstrated in a comparison to state-of-the-art methods.

Voronoi coverage control is a particular problem of importance in the area of multi-robot systems, which considers a network of multiple autonomous robots, tasked with optimally covering a large area. This is a common task for fleets of fixed-wing Unmanned Aerial Vehicles (UAVs), which are described in this work by a unicycle model with constant forward-speed constraints. We develop event-based control/communication algorithms to relax the resource requirements on wireless communication and control actuators, an important feature for battery-driven or otherwise energy-constrained systems. To overcome the drawback that the event-triggered algorithm requires continuous measurement of system states, we propose a self-triggered algorithm to estimate the next triggering time. Hardware experiments illustrate the theoretical results.

We introduce a distributed control method for coordinating multiple vehicles in the framework of an automated valet parking (AVP) system. The control functionality is distributed between an infrastructure server, called parking area management (PAM) system, and local autonomous vehicle control units. Via a vehicle-to-infrastructure (V2I) communication interface, model predictive control (MPC) decisions of the vehicles are shared with the coordination unit in the PAM. This unit in turn computes a coupling feedback which is shared with the vehicles. The control system is integrated in an automated test-system to cope with the high test requirements and short development cycles of highly automated systems. Evaluations conducted with the test-system show the functionality of the proposed distributed control method for multi-vehicle coordination. Results indicate safe coordination, and an efficiency increase compared to an uncoordinated method in an AVP simulation environment.

Age of information (AoI) measures information freshness at the receiver. AoI may provide insights into quality of service in communication systems. For this reason, it has been used as a cross-layer metric for wireless communication protocols. In this work, we employ AoI to calculate penalty functions for a centralized resource scheduling problem. We consider a single wireless link shared by multiple, heterogeneous control systems where each sub-system has a time-varying packet loss probability. Sub-systems are competing for network resources to improve the accuracy of their remote estimation process. In order to cope with the dynamically changing conditions of the wireless link, we define a finite horizon age-penalty minimization problem and propose a scheduler that takes optimal decisions by looking H slots into the future. The proposed algorithm has a worst-case complexity that grows exponentially with H. However, by narrowing down our search space within the constrained set of actions, we are able to decrease the complexity significantly without losing optimality. On the contrary, we show by simulations that the benefit of increasing H w.r.t. remote state estimation performance diminishes after a certain H value.

Cooperative manipulation tasks can be divided into the subtasks of object trajectory tracking and grasp maintenance. Both subtasks typically pose individual requirements on the underlying control objective in terms of accuracy and robustness to disturbances. In this article, we propose a novel distributed impedance control scheme, resulting in a flexible control design to meet the individual-potentially conflicting-impedance goals of the respective subtasks. In order to achieve more efficient use of the communication resource, we propose an event-triggered strategy for the communication between the robotic agents. Robust, stability also in interaction with unknown objects and unknown environments, is guaranteed via passivity-based control design. Simulations and experimental results show that, with the proposed communication strategy, the number of transmissions between agents can be significantly reduced while maintaining the flexibility introduced by the distributed controller.

This article investigates a fully distributed cooperation scheme for networked mobile manipulators. To achieve cooperative task allocation in a distributed way, an adaptation-based estimation law is established for each robotic agent to estimate the desired local trajectory. In addition, wrench synthesis is analyzed in detail to lay a solid foundation for tight cooperation tasks. Together with the estimated task, a set of distributed adaptive (DA) controllers is proposed to achieve motion synchronization of the mobile manipulator ensemble over a directed graph with a spanning tree irrespective of the kinematic and dynamic uncertainties in both the mobile manipulators and the tightly grasped object. The controlled synchronization alleviates the performance degradation caused by the estimation/tracking discrepancy during the transient phase. The proposed scheme requires no persistent excitation condition and avoids the use of noisy Cartesian-space velocities. Furthermore, it is independent from the object's center of mass by employing formation-based task allocation and a task-oriented strategy. These attractive attributes facilitate the practical application of the scheme. It is theoretically proven that convergence of the cooperative task tracking error is guaranteed. Simulation results, as well as manipulation experiments with three mobile manipulators involved, validate the efficacy and demonstrate the expected performance of the proposed scheme.

Data-driven models are subject to model errors due to limited and noisy training data. Key to the application of such models in safety-critical domains is the quantification of their model error. Gaussian processes provide such a measure and uniform error bounds have been derived, which allow safe control based on these models. However, existing error bounds require restrictive assumptions. In this paper, we employ the Gaussian process distribution and continuity arguments to derive a novel uniform error bound under weaker assumptions. Furthermore, we demonstrate how this distribution can be used to derive probabilistic Lipschitz constants and analyze the asymptotic behavior of our bound. Finally, we derive safety conditions for the control of unknown dynamical systems based on Gaussian process models and evaluate them in simulations of a robotic manipulator.

Teleoperation of multi-robot systems, e.g. dual manipulators, in cooperative manipulation tasks requires haptic feedback of multi-contact interaction forces. Classical haptic devices restrict the workspace of the human operator and provide only one contact point. An alternative solution is to enable the operator to command the robot system via free-hand motions which extends the workspace of the human. In such a setting, a multi-contact haptic feedback may be provided to the human through multiple wearable haptic devices, e.g. fingertip devices that display forces on the human fingertips. In this paper we evaluate the benefit of using wearable haptic fingertip devices to interact with a bimanual robot setup in a pick-and-place manipulation task. We show that haptic feedback through wearable devices improves task performance compared to the base condition of no haptic feedback. Therefore, wearable haptic devices are a promising interface for guidance of multi-robot manipulation systems.

In this paper, we consider a dynamical system whose trajectory is a result of minimizing a multiphase cost function. The multiphase cost function is assumed to be a weighted sum of specified features (or basis functions) with phase-dependent weights that switch at some unknown phase transition points. A new inverse optimal control approach for recovering the cost weights of each phase and estimating the phase transition points is proposed. The key idea is to use a length-adapted window moving along the observed trajectory, where the window length is determined by finding the minimal observation length that suffices for a successful cost weight recovery. The effectiveness of the proposed method is first evaluated on a simulated robot arm, and then, demonstrated on a dataset of human participants performing a series of squatting tasks. The results demonstrate that the proposed method reliably retrieves the cost function of each phase and segments each phase of motion from the trajectory with a segmentation accuracy above 90%.

This video features RAMCIP, a new service robot developed to provide proactive and discreet assistance to elderly with Mild Cognitive Impairments (MCI), supporting their daily activities at home. Starting with a thorough analysis of needs and requirements of the target population, the RAMCIP robot was developed as an integrated ensemble of advanced H/W and S/W components, realizing the robot skills of perception, cognition, safe navigation, grasping, manipulation, and human-robot communication, ample to operate in real, rather challenging domestic environments. The RAMCIP use-cases include proactive assistance provision to user's cooking, eating and medication activities, through discreet user monitoring and robot interventions by reminders and robotic manipulations., RAMCIP can bring the medicine, recognize fallen objects and electric appliance that has been forgotten turned on. It also recognizes the user walking in low-light conditions and turns on the light, as well as detects cases of emergency such as a fall. The robot provides also the user with cognitive training games and stimulates the user to contact with relatives through video-calls. Pilot trials of the RAMCIP robot have been performed in real homes of more than ten different users, in Barcelona, Spain; the video at hand exhibits the robot performing the target use cases.

As the application of probabilistic models in robotic applications increases, a systematic robot control approach considering the effects of uncertainty becomes indispensable. Inspired by human sensorimotor findings, in this paper, we study the stochastic optimal control problem with high-order cost statistics in order to synthesize uncertainty-dependent actions in robotic scenarios with multiple uncertainty sources. We present locally optimal risk-sensitive and cost-cumulant solutions for settings with nonlinear dynamics, multiple additive uncertainty sources, and nonquadratic costs. The influence of each uncertainty source on the cost can be individually parameterized offering additional flexibility in the control design. We further analyze the case in which the static uncertain parameters are involved. The simulations of several linear and nonlinear settings with nonquadratic costs and an experiment on a real robotic platform validate our approach and illustrate its peculiarities.

The planning and execution of real-world robotic tasks largely depends on the ability to generate feasible motions online in response to changing environment conditions or goals. A spline deformation method is able to modify a given trajectory so that it matches the new boundary conditions, e.g., on positions, velocities, accelerations, etc. At the same time, the deformed motion preserves velocity, acceleration, jerk, or higher derivatives of motion profile of the precalculated trajectory. The deformed motion possessing such properties can be expressed by translation of original trajectory and spline interpolation. This spline decomposition considerably reduces the computational complexity and allows real-time execution. Formal feasibility guarantees are provided for the deformed trajectory and for the resulting torques. These guarantees are based on the special properties of Bernstein polynomials used for the deformation and on the structure of the chosen computed-torque control scheme. The approach is experimentally evaluated in a number of planar volleyball experiments using 3 degree-of-freedom robots and human participants.

While convolutional neural networks (CNNs) have been successfully applied to many challenging classification applications, they typically require large datasets for training. When the availability of labeled data is limited, data augmentation is a critical preprocessing step for CNNs. However, data augmentation for wearable sensor data has not been deeply investigated yet.

A data-driven identification of dynamical systems requiring only minimal prior knowledge is promising whenever no analytically derived model structure is available, e.g., from first principles in physics. However, meta-knowledge on the system’s behavior is often given and should be exploited: Stability as fundamental property is essential when the model is used for controller design or movement generation. Therefore, this paper proposes a framework for learning stable stochastic systems from data. We focus on identifying a state-dependent coefficient form of the nonlinear stochastic model which is globally asymptotically stable according to probabilistic Lyapunov methods. We compare our approach to other state of the art methods on real-world datasets in terms of flexibility and stability.

In this paper we consider the problem in which the human commands the overall behavior of a robot team while the robots are controlled to comply with formation constraints. Such human-robot team interaction is challenging in terms of system complexity and control synthesis. Port-Hamiltonian framework is suitable for modeling the interconnected systems. In this paper we model the robotic team, cooperatively manipulating an object, as a constrained port-Hamiltonian system. Furthermore, we propose a passivity-based control approach in the port-Hamiltonian framework for the cooperative manipulation system guided by the human. The control mechanism is based on the energy shaping for achieving a desired behavior of the formation and its preservation. An energy tank in the cascade is introduced to guarantee passivity of the system commanded by the human and safe interaction with humans in the robot environment. We validate the proposed approach with simulation and experiments.

Knowing accurately the dynamic parameters of a manipulated object is required for common coordination strategies in physical human-robot interaction. Bias in object dynamics results in inaccurately calculated robot wrenches, which may disturb the human during interaction and bias the recognition of the human motion intention. This paper presents an identification strategy of object dynamics for physical human-robot interaction, which allows the tracking of desired human motion and inducing the motions necessary for parameter identification. The estimation of object dynamics is performed online and the estimator minimizes the least square error between the measured and estimated wrenches acting on the object. Identification-relevant motions are derived by analyzing the persistence of excitation condition, necessary for estimation convergence. Such motions are projected in the null space of the partial grasp matrix, relating the human and the robot redundant motion directions, to avoid disturbance of the human desired motion. The approach is evaluated in a physical human-robot object manipulation scenario.

Programming by Demonstration allows to transfer skills from human demonstrators to robotic systems by observation and reproduction. One aspect that is often overlooked is that humans show different trajectories over multiple demonstrations for the same task. Observed movements may be more precise in some phases and more diverse in others. It is well-known that the variability of the execution carries important information about the task. Therefore, we propose a Bayesian approach to model uncertainties from training data and to infer them in regions with sparse information. The approach is validated in simulation, where it shows higher precision than existing methods, and a robotic experiment with variance based impedance adaptation.

Robot teams require planning and adaptive capabilities in order to perform cooperative manipulation tasks in dynamic or unstructured environments. Since these capabilities are inherent to humans, it is suitable to consider human-robot team teleoperation for cooperative manipulation where a single human collaborates with the robot team. In this paper, we present a subtask-based control approach which enables a simultaneous execution of two subtasks by the robot team, interacting with the object: trajectory tracking and formation preservation. Control inputs for both subtasks are provided by the human operator. The commands are projected onto the spaces of subtasks using a command mapping strategy. Analogously, measured interacting forces are projected onto the space of feedback signals, provided to the human via wearable fingertip haptic devices through a feedback mapping strategy. Experimental results validate the proposed approach.

In human-robot interaction, it is essential to ensure that the robot poses no threat to the human. Especially in applications that require close or physical interaction, e.g., collaborative manufacturing or rehabilitation, the danger emanating from the robot has to be minimized. Control schemes introducing virtual constraints have proven valuable in this context since they allow us to define a safe zone to move in without endangering the human. Combining the different requirements on the control scheme, such as real-time capability, stability, and reliability, in the presence of external disturbances and dynamic limits, however, turns out to be challenging. In this paper, we present a novel control scheme for human-robot interaction, which enforces dynamic constraints even in the presence of external forces. Based on an analytic constraint description and a feedback linearization of the system dynamics, a safe set of states is determined, which is then rendered controlled positively invariant, thus keeping the system in a safe configuration. The controlled system is analyzed with respect to invariance and boundedness with the results being illustrated in a full-scale experiment.

Robots using distributed sensors in smart environments and smart infrastructure systems such as traffic and power systems are examples of networked cyber-physical systems where communication and/or computational resources are constrained. The scientific challenge is to design scheduling and control schemes taking into account such resource constraints and to preferably include fair resource sharing mechanisms among different control applications. In this talk we present a novel framework for the optimal co-design of scheduling and control for networked systems with resource constraints. In particular we consider multiple control loops, which transmit their measurements over a shared communication channel. Only a limited number of those control loops may close their feedback loop at a time. As a result the dynamics of the individual control loops are coupled through the resource constraint. The scientific question is, when a control loop should schedule the transmission of a measurement and what is the appropriate control law. We approach the problem from an optimality point of view with the scheduling and control policies being the optimization variables. We derive an efficient and tractable decomposition, which allows a distributed solution for control and scheduling decisions coordinated by a price-based mechanism. It turns out that an event-triggered control scheme is optimal and that certainty equivalence holds. In fact, our scheme exploits the adaptation ability of event-triggered control in terms of communication traffic elasticity. Furthermore, we provide stability results linking the resource constraints with the system dynamics.

Control over shared communication networks is a key challenge in design and analysis of cyber-physical systems. The quality of control in such systems might be degraded due to the congestion while accessing the scarce communication resources. In this paper, we consider a multiple-loop networked control system (NCS), where all control loops share a communication network. Medium Access Control (MAC) is performed in contention-based fashion using a multi-channel slotted ALOHA protocol, where each control loop decides locally whether to attempt a transmission based on some error thresholds. We further introduce a local event-based resource-aware scheduling design with an adaptive choice of the error thresholds for a transmission. This leads to a hybrid channel access mechanism where the control loops are deterministically categorized into two sets of eligible and ineligible sub-systems for transmission in an event-based fashion, before a random process to select the available channels. In addition, employing the introduced policy, we show the stability of the resulting NCS in terms of Lyapunov stability in probability. We illustrate numerically the efficiency of our proposed approach in terms of reducing the average networked-induced error variance, and show the superiority of the adaptive event-based scheduler compared to the scheduling design with non-adaptive thresholds.

For seamless physical human-robot interaction (pHRI), estimating human intention is essential. Most system identification approaches to pHRI model the human as a black box without prior assumptions about the underlying behavioral structure. However, integrating a priori knowledge about behavioral characteristics of the human provides superior prediction performance. In this work we present a novel method for human behavior prediction during physical interaction that incorporates an empirically supported human motor control model. The arm dynamics of the human are modeled as a mechanical impedance that follows a latent desired trajectory. We adopt a Bayesian perspective setting Gaussian Process (GP) priors on impedance parameters and the desired trajectory, which allows regression about human behavior from observed trajectories and interaction forces. The proposed impedance-based GP model is validated in simulation and in an experiment with human participants to demonstrate its prediction performance and generalization capability.

Dynamic Movement Primitives (DMPs) represent stable goal-directed or periodic movements, which are learned from observations or demonstrations. They rely on proper function approximators, which are sufficiently flexible to represent arbitrary movements but also ensure goal convergence in point-to-point motions. This work shows that Gaussian Processes (GPs) are suitable as a regressor for learning movements with DMPs ensuring stability. In addition, GPs provide a measure for the uncertainty about the current movement, which we exploit by proposing a new cooperation scheme for DMPs: For better reproduction of demonstrations, we follow the intuition, that individuals with more knowledge lead towards the goal, while others follow and focus on cooperation. Along with simulation results, we validate the presented methods in a robotic cooperative object manipulation task.

Many robotic applications, especially if humans are involved, require the robot to adhere to certain joint, workspace, velocity or force limits while simultaneously executing a task. In this paper, we introduce a control structure, which merges an arbitrary desired robot behavior with given constraints. Using a quadratic program (QP), control barrier functions (CBFs) are combined with an arbitrary nominal control law, which determines the desired behavior. The CBFs enforce the constraints, overruling nominal control whenever necessary. We show that the concept is applicable with arbitrary numbers of constraints and any nominal control law. In order to illustrate the capabilities of the approach, the control scheme is applied to an anthropomorphic manipulator, which is constrained by static as well as moving constraints.

Ongoing technological advances in the areas of computation, sensing, and mechatronics enable robotic-based systems to interact with humans in the real world. To succeed against a human in a competitive scenario, a robot must anticipate the human behavior and include it in its own planning framework. Then it can predict the next human move and counter it accordingly, thus not only achieving overall better performance but also systematically exploiting the opponent's weak spots. Pool is used as a representative scenario to derive a model-based planning and control framework where not only the physics of the environment but also a model of the opponent is considered. By representing the game of pool as a Markov decision process and incorporating a model of the human decision-making based on studies, an optimized policy is derived. This enables the robot to include the opponent's typical game style into its tactical considerations when planning a stroke. The results are validated in simulations and real-life experiments with an anthropomorphic robot playing pool against a human.

Teleoperation in extreme environments may suffer from communication delay and packet loss during the transmission of command signals and sensory feedback. This study investigates whether online control of communication time delay by using quality of service (QoS) techniques can improve operator task performance in a virtual teleoperated collision avoidance task. We first introduce the framework of predictive communication quality control based on a dynamic performance model of a human handling the teleoperation system. We then apply the framework to a virtual collision avoidance scenario and evaluate it with two behavioral studies. Study 1 identifies that prolonging time delay significantly increases the frequency of collisions and completion time. We develop a model for predicting the probability of the operator causing collisions with the wall and fit its parameters with the experimental data. In Study 2, we compare the completion time and the number of collisions with and without the predictive QoS control. It is shown that the predictive QoS control is capable of reducing the number of collisions, but it does not affect task completion time. The prediction model and empirical validation provide a successful proof of concept for a human-centered system design, in which the dynamic model of the operator is the center of the control architecture.

Efficient coordination of a multirobot team is the key challenge in robotic application domains such as manufacturing, construction, and service robotics. In cooperative manipulation tasks, the system dynamics result from the complex interaction of several manipulators handling a common object. A comprehensive model is indispensable for a sophisticated model-based control design. An open problem is the modeling and analysis of the overall system dynamics including the manipulators' interaction wrenches. Based on the apparent end-effector dynamics in task space, in this paper we focus on the characterization of the interaction effects when manipulating a common object. We note the central role of the imposed kinematic constraints for the emerging system dynamics, their significance for the manipulator coordination in terms of control design, and the analysis of internal wrenches applied to the object. We derive fundamental properties of the cooperative manipulator system relevant to the manipulation task such as the apparent impedance with respect to external disturbances. An experimental study is conducted with two cooperating anthropomorphic manipulators supporting the relevance of our findings.

While research has demonstrated how vibrotactile devices can be effectively used to guide human behavior, efficient mappings of vibration patterns for spatial guidance in time-critical dynamical tasks have not yet been understood. In this paper, we contrast two types of action-dependent, haptic stimulus designs to demonstrate the different effects of vibrotactile feedback on the human control performance. A wireless bracelet is used to provide patterns of vibrotactile stimuli in real-time, representing either optimal hand velocity or acceleration for the stabilization of an inverted pendulum. The optimal control behavior is supplied by a linear quadratic regulator. The analyses of the participants' stabilization and learning behavior revealed a significant improvement caused by the additional velocity-dependent feedback. The results are consistent with previous research, which indicates that the human sensory-motor system is generally more sensitive to velocity than acceleration information. In summary, the present paper suggests how human-centric vibrotactile stimuli should be designed and how they can be effectively transmitted to the human user for time-critical behavioral guidance.

This paper discusses an online dynamic motion generation scheme for nonprehensile object manipulation by using a set of predefined motions and a trajectory deformation algorithm capable of incorporating positional and velocity boundary constraints. By creating optimal trajectories offline and deforming them online, computational complexity during execution is reduced considerably. As tight convex hulls of the deformed trajectories can be found, possible obstacles or workspace boundaries can be circumnavigated precisely without collision. The approach is verified through experiments on an inclined planar air-table for volleyball scenario using two 3-DoF robots.

In cooperative manipulation tasks, load allocation is a crucial step in order to solve the intrinsic redundancy of the system. The desired wrench needs to be suitably distributed between the end-effectors to implement the desired motion of the manipulated object. In this framework, both the grasp kinematics and the individual capacities of each manipulator provide relevant constraints. On one hand, the end effector wrenches act on the object via the grasp geometry. On the other hand, the individual admissible payload further depends on the current configuration of the robots. In this paper we focus on a heterogeneous cooperative manipulation setting and we design a proper allocation strategy to distribute the desired object wrench, considering both constant and time-varying constraints for the load distribution. The relevance of our findings is illustrated by means of an experimental study involving two anthropomorphic robots manipulating a common object.

Knowledge of the human grasp pose is crucial in common control schemes for human-robot object manipulation tasks. Biased estimates of the grasp pose cause undesired interaction wrenches on the human partner, which disturbs the interaction and the recognition of motion intention. A use of wearable motion sensors for tracking the human motion facilitates the grasp pose estimation without a global sensing system. This paper presents an approach for estimating an unknown grasp pose of the human using wearable motion sensors while minimizing undesired interaction wrenches applied to the human. A condition necessary for convergence of the estimator together with appropriate robot motion strategies are provided. Estimation of relative orientation and displacement is performed online and based on minimizing the error in the least-square sense. The estimation process does not rely on a global sensing system and it considers only the measurements of the velocity and acceleration of the cooperating partners in their respective local frames. The approach is experimentally evaluated in a physical human-robot interaction scenario.

In human-robot interaction (HRI) and especially in close or physical interaction, it is essential to ensure the human's safety. This is achieved by introducing virtual constraints defining a region, in which the robot is allowed to move safely. These safety regions may change over time during human-robot interaction, which may be either due to human motion or changed environmental conditions. In consequence it is important for the applied control scheme to handle dynamic boundaries. This work proposes an invariance-based control approach, which enforces adherence to boundaries with dynamic parameters. We extend the invariance control approach, which provides a computationally efficient and systematic method for defining constraints on system states and outputs, such that it handles the constraint dynamics. Stability and invariance properties are analyzed and validated in an experimental evaluation on a 7-DoF anthropomorphic manipulator.

As the application of probabilistic models in robotic applications increases, the necessity of a systematic robot-control method that considers the effects of multiple uncertainty sources becomes more evident. Motivated by human sensorimotor findings, in this work we study the stochastic locally optimal feedback control problem with high-order cost statistics where dynamics have multiple additive noise sources and cost variability produced by each uncertainty source is evaluated marginally. We present risk-sensitive and cost-cumulant solutions for this problem for non-linear dynamics and non-quadratic costs. Locally optimal solutions are found by iteratively performing a linear quadratic approximation around a nominal trajectory, solving the local problem and updating the trajectory until convergence. Simulation results of a point mass robot and a two-link manipulator validate the applicability of the proposed approach and illustrate its peculiarities.

The interaction of a single human with a team of cooperative robots, which collaboratively manipulate an object, poses a great challenge by means of the numerous possibilities of issuing commands to the team or providing appropriate feedback to the human. In this paper we propose a formation-based approach in order to avoid deformations of the object and to virtually couple the human to the formation. Here the human can be interpreted as a leader in a leader-follower formation with the robotic manipulators being the followers. The results of a controllability analysis in such a leader-follower formation suggest that it is beneficial to measure the state of the human (leader) by all physically cooperating manipulators (followers). The proposed approach is evaluated in a full-scale multi-robot cooperative manipulation experiment with humans.

The load distribution strategy in cooperative manipulation tasks allocates suitable force and torque setpoints to an ensemble of manipulators in order to implement a desired action on the manipulated object. Due to the manipulator redundancy, the load distribution computed by means of a generalized inverse of the grasp matrix is not uniquely determined. Controversial results on the nonsqueezing property of specific load distributions exist in the literature. In this paper, we propose a new paradigm for the analysis of internal wrenches based on the kinematic constraints imposed to the manipulator ensemble. We unify previous results by showing that there exists no unique nonsqueezing load distribution and illustrate the consequences of our findings by means of several examples. In particular, the presented results provide a new perspective on the decomposition of interaction forces into internal and external components as required for cooperative multimanipulator control schemes.

Intuitive and effective physical assistance is an essential requirement for robots sharing their workspace with humans. Application domains reach from manufacturing and service robotics via rehabilitation and mobility aids to education and training. In this context, assistance based on human behavior anticipation has shown superior performance in terms of human effort minimization. However, when a robot's expectations mismatch a human intentions, undesired interaction forces appear incurring safety risks and discomfort. Human behavior prediction is, therefore, a crucial issue: It enables effective anticipation but potentially produces disagreements when prediction errors occur. In this paper, we present a novel control scheme for anticipatory haptic assistance where robot behavior adapts to prediction uncertainty. Following a data-driven stochastic modeling approach, robot assistance is synthesized solving a risk-sensitive optimal control problem, where the cost function and plant dynamics are affected by model uncertainty. The proposed approach is objectively and subjectively evaluated in an experiment with human users. Results indicate that our method outperforms other assistive control approaches in terms of perceived helpfulness and human effort minimization.

The workspace and performance of a humanoid robot is decisively influenced by the design of its torso. The joints or spinal discs are usually the weak points due to the high stress they are exposed to, e. g. when lifting heavy objects. One way to circumvent the necessity of large motors is to use parallel mechanisms to optimize the distribution of loads. Here, we analyze the workspace of the humanoid robot Rollin' Justin of the German Aerospace Center (DLR) w. r. t. the constraints imposed by kinematic coupling of torso joints via tendons. The results of the analysis can be used for planning and reactive control to efficiently exploit the torso performance capabilities of the robotic system. As an application, we design a potential field based controller to avoid violating these constraints and implement it on the real robot.

Gaussian Processes (GPs) are gaining increasing popularity due to their expressive power for learning the dynamics of non-linear time series data, e.g. for human motion prediction. However, so far they are restricted to Euclidean space: input data such as position and velocity need to be Euclidean. In this paper, we examine GPs over time series of 6D rigid body motions including large rotations. As the use of Euler angles with large rotations results in inaccurate predictions, we present an extension of the valid input data to quaternions H and dual quaternions HD. The quality of a GP prediction over unit quaternions is compared with GP prediction over Euler angles. The results are evaluated based on experimental data from a quadrotor and in a learning task of a collision free 6D motion trajectory incorporating large rotations based on artificial data from a motion planner.

This paper presents a novel robot pose measure for human movement imitation based entirely on the Euclidean distance information between any two links of a robot and any link and object in the robot's environment in a Cartesian task space. A Hidden Markov Model is used to encode the spatio-temporal information of multiple demonstrations. In combination with Gaussian Mixture Regression for extracting the important task properties, feasible full-body motion adaption can be achieved. The method is suited for use with a humanoid robot by considering additional constraints like balance control and collision avoidance. In order to tackle modeling errors occurring due to the human movement demonstration and the robotic reproduction, a manipulability based weighting scheme is proposed. Complexity reduction of the otherwise redundant pose measure is performed based upon a mechanical analogy of an interconnected spring system. Experiments are conducted using a HRP-4 robot and display the applicability of the presented methods for robotic full-body motion imitation tasks.

Cooperative manipulation, where several robots jointly manipulate an object from an initial configuration to a final configuration while preserving the robot formation, poses a great challenge in robotics. Here, we treat the problem of designing motion primitives for cooperative manipulation such that the robots move in formation and are robust with respect to external disturbances. Individual robot trajectories are generated by Dynamic Movement Primitives (DMPs) and coupled by a formation control approach enabling the DMP-trajectories to preserve a given formation while performing the manipulation. The proposed control scheme achieves an increased adaptability under external disturbances. The approach is evaluated in a full-scale experiment with two prototypical cooperative manipulation and synchronized motion tasks.

This paper presents an incremental sampling-based approach for trajectory imitation in cluttered environments using the RRT* algorithm. Inspired by the discrete Laplace-Beltrami operator the underlying distance metric is based upon the difference from a reference trajectory through a quadratic distance term incorporating velocity and acceleration deviations along the trajectory. Mathematically-backed approximations in combination with a task-space bias make it possible to use standard nearest neighbor methods in task space when expanding the RRT*-tree. It is shown that metric-consistent biases considerably increase the convergence speed. The proposed approach is validated in simulations in a 2D environment and in experiments using a HRP-4 humanoid robot.

Cooperative dynamic object manipulation can extend the manipulation capabilities of robot-robot and human-robot teams. In order to be able to inject energy into various suspended objects of unknown parameters, in this paper we propose an adaptive controller which combines reinforcement learning with energy based swing-up control. The proposed controller is successfully verified in a single robot and human-robot experimental setup for different types of suspended objects.

This paper investigates goal-directed cooperative object swinging as a novel physical human-robot interaction scenario. We develop an energy-based control concept, which enables a robot to cooperate with a human in a goal-directed swing-up task. The robot can be assigned to be a leader or an actively contributing follower. We conduct a virtual reality experiment to compare effort sharing and performance of mixed human-human and human-robot dyads. The leader and the follower controllers yield similar results compared to their human standard.

Trajectory generation for active physical assistance to humans in cooperative haptic tasks gains increasing interest in recent literature. Planning-based approaches represent one class of trajectory synthesis methods for active robotic partners. To overcome the limitations of kinematic planning algorithms in dynamic tasks, we propose a three-step approach to the synthesis of trajectories under the principle of least action. This is motivated by neuroscientific findings on human effort minimization in motor tasks. A trajectory is generated by optimized sequencing of optimal motion primitives. The benefits of the proposed method for physical human-robot cooperation are demonstrated in human user studies in a 2D cooperative transport task in a virtual maze.

It is well-known that physical robotic assistance to humans is significantly enhanced by including human behavior anticipation into robot planning and control. The challenge arises when the human goal/plan is uncertain or unknown to the robot. In this paper we propose a novel control scheme which dynamically selects between a model-based and a model-free strategy depending on the level of disagreement between the human and the robot. The disagreement is measured in terms of the interaction force. A task specific model-based controller is selected when the human's motion intention coincides with the robot's goal. A model-free control scheme based on the human force as motion prediction source is selected in case of disagreement and when the human goal/plan is unknown. The benefits of this approach are demonstrated in a human user study on human-robot cooperative object transport through a 2D maze in virtual reality.

Manipulation tasks are a great challenge for robots due to the uncertainty arising from unstructured environments. In this paper we propose a novel control scheme for contact tasks based on risk-sensitive optimal feedback control. It provides a systematic approach to adjust the trade-off between motion and force control under uncertainty. Following a previously acquired task model, the proposed approach provides both a variable stiffness solution and a motion reference adaptation. This control scheme achieves increased adaptability under previously unseen environmental variability. An implementation on a robotic manipulator validates the applicability and adaptability of the proposed control approach in two different manipulation tasks.

Multi-robot cooperative manipulation of a common object requires precise kinematic coordination of the attached end effectors in order to avoid excessive forces on the object and the manipulators. A manipulation task is considered successful if the desired object motion and forces are tracked accurately. In this paper we present a systematic analysis on the effect of uncertain kinematic parameters on the tracking behavior in a planar manipulation task. An adaptive control scheme is proposed, which achieves the desired control goal asymptotically. The presented scheme employs the current force/motion data of the attached end effectors without relying on a common reference frame. The algorithm is applicable to common manipulator types with wrist-mounted force/torque sensors and implementable in real-time. The performance of the proposed control scheme is evaluated experimentally with two 7DoF manipulators who cooperatively manipulate an object of uncertain length.

Cooperative manipulation in robotic teams likely results in an increased manipulation performance due to complementary sensing and actuation capabilities or increased redundancy. However, a precise coordination of the involved manipulators is required in order to avoid undesired stress on the manipulated object. Extending the workspace of the robots by means of mobile platforms greatly enlarges the potential task spectrum but simultaneously poses new challenges for example in terms of increased kinematic errors. In this paper we show how kinematic errors in the closed kinematic chain originating from uncertainties in the geometry of object and manipulators limit the cooperative task performance. We extend an impedance-based coordination control scheme towards mobile multi-robot manipulation to limit undesired internal forces in the presence of kinematic uncertainties. Furthermore, we employ a task-space decoupling approach to reduce the impact of disturbances at the mobile platforms on the end effectors. The presented control scheme for cooperative, mobile dualarm manipulation is applicable in real-time and suitable for a team of heterogeneous manipulators. We evaluate the presented architecture by means of a large-scale experiment with four 7DoF manipulators on two mobile platforms.

Cooperative manipulation, where several robots collaboratively transport an object, poses a great challenge in robotics. In order to avoid object deformations in cooperative manipulation, formation rigidity of the robots is desired. This work proposes a novel linear state feedback controller that combines both optimal goal regulation and a relaxed form of the formation rigidity constraint, exploiting an underlying distributed impedance control scheme. Since the presented control design problem is in a biquadratic LQR-like form, we present an iterative design algorithm to compute the controller. As an intermediate result, an approximated state-space model of an interconnected robot system is derived. The controller design approach is evaluated in a full-scale multi-robot experiment.

Adjusting to new situations by changing the shape of a prerecorded trajectory is an important aspect for robot manipulation in a constrained environment. For being recognized as a distinctive trajectory, the goal of any trajectory modification is to keep local and global properties as similar as possible compared to the reference trajectory. This paper presents a framework that can alter the shape of a trajectory by defining the position of a set of sampling points while maintaining local properties in a least-squares manner. The method consists of a three-staged approach first modifying the global shape of the trajectory and subsequently taking local features into account. Inspired by mesh processing used for 3D surface editing, differential coordinates based on the discretized Laplacian operator are used for measuring and maintaining local trajectory properties when deforming the trajectory. Last, a post-processing step based on a relaxed “as-rigid-as-possible” principle allows local deformations and length modifications of the trajectory for a better tradeoff between preserving local and global properties. Experiments verifying the applicability of the proposed algorithm are conducted using a 7-DoF anthropomorphic arm following a previously recorded and modified trajectory.

While human behavior prediction can increase the capability of a robotic partner to generate anticipatory behavior during physical human robot interaction (pHRI), predictions in uncertain situations can lead to large disturbances for the human if they do not match the human intentions. In this paper we present a novel control concept in which the assistive control parameters are adapted to the uncertainty in the sense that a the robot takes a more or less active role depending on its confidence in the human behavior prediction. The approach is based on risk-sensitive optimal feedback control. The human behavior is modeled using probabilistic learning methods and any unexpected disturbance is considered as a source of noise. The proposed approach is validated in situations with different uncertainties, process noise and risk-sensitivities in a tow- Degree-of-Freedom virtual reality experiment.

In this video we present a teleoperation system which is capable of solving complex tasks in human-sized wide area environments. The system consists of two mobile teleoperators controlled by two operators, and offers haptic, visual, and auditory feedback. The task examined here, consists of repairing a robot by removing a computer and replacing a defective hard-drive. To cope with the complexity of such a task, we go beyond classical teleoperation by integrating several advanced software algorithms into the system.

For safety in physical human-robot interaction (pHRI) the robot motion must be restricted to an admissible (safe) region. In this work, we propose a systematic approach to guarantee the satisfaction of virtual workspace constraints in 6D for arbitrary manipulator dynamics based on an extended invariance control concept. Invariance control yields a computationally efficient method to render multiple virtual nonlinear workspace boundaries. In order to make the scheme suitable for pHRI we present an approach to reduce chattering by explicitly considering the discrete-time Euler solver output. Orientation constraints are unambiguously represented as unit quaternions. The theoretical results are successfully validated in simulation and experiments on a 7-DoF anthropomorphic manipulator.

Humans exhibit exceptional skills in using tools and manipulating objects of their environment by skillfully controlling exerted force and arm impedance. One of the basic components of this mechanism is the generation of internal models which associate kinematic variables with applied force. On the other hand, making robots capable of skillfully using tools and adapting their motor behavior to new environmental conditions is rather complex. In the present paper, we investigate learning of force control policies for robotic sculpting given multiple task demonstrations. These policies express the relationship between constrained motions and exerted force and are learned in Cartesian space where the coupling of dynamics between different directions of motion is also taken into account. In addition, a novel algorithm is proposed to generalize these policies to new motion tasks, executed in a sufficiently homogeneous environment, same with that in demonstrations, but in presence of new motion-dependent external forces. To this aim, a differential calculus approach is proposed where not only the mapping from motion to force but also from difference in motion to difference in force is learned to generalize the policies to new contexts. This is achieved by learning apart from a set of policy parameters, some newly introduced quantities, so called weight differentials, which express the rate of change of the policy parameters. The proposed approach is validated in simple real-world sculpting experiments by using a two degrees-of-freedom haptic device.

Goal-directed physical assistance to the human is one of the most challenging problems in the area of human-robot interaction. Planning and learning from demonstration represent two conceptually different approaches to achieve goal-directed behavior. Here we examine the properties of a planning-based and a learning-based approach in the context of physical robotic assistance for the prototypical task of cooperative object maneuvering. In order to exploit the complementary strengths of planning and learning-based approaches we derive three novel synergy strategies. The algorithms are experimentally evaluated in a human user study in a planar virtual-reality scenario and in a proof-of-concept study with a human-sized mobile robot with two 7DoF arms. The results show that combinations of planning and learning algorithms are superior over the individual approaches.

Proactive physical robotic assistance in the presence of human prediction uncertainty is a very challenging control problem. In this paper we propose a risk-sensitive optimal feedback controller for physical assistance that autonomously adapts the robot's behavior even during unknown situations. Using a probabilistic model to represent the cooperative task execution behavior and modeling the human as a source of process noise in the system, the proposed assistive controller proactively contributes to the task anticipating the human motion. Estimating online the current level of disagreement and prediction uncertainty, the assistive controller consequently calculates the optimal task contribution providing higher adaptability. A psychological evaluation compares different assistive control strategies in a virtual scenario using a two-Degree-of-Freedom haptic experimental setup. Results show that considering the current level of disagreement enhances the performance of the controller in terms of helpfulness and human effort minimization.

This paper presents a novel descriptor for trajectory classification in n dimensions, which is invariant with respect to scaling and rigid transformation. Using a hierarchical approach, the descriptor is able to capture both local and global features of the trajectory. The algorithm iteratively splits up every trajectory into smaller trajectory segments resulting in a binary tree. Inspired by the Frenet-Serret formulas, a projection onto a lower dimensional subspace is performed for every trajectory segment, providing a characteristic description of every trajectory. The subspace projection acts as a pseudo-curvature measure in every dimension. Successful applicability is shown through classification experiments in three and six dimensions using an RGB-D camera. For comparison with other algorithms, the Australian Sign Language dataset is also used for classification, showing a superior classification rate.

In comparison to industrial settings with structured environments, the operation of autonomous robots in unstructured and uncertain environments is more challenging. This video presents a generic control and system architecture ARCADE, applicable for real-time robot control in complex task situations. Several methods to cope with uncertainties are demonstrated with the example task of changing tires on a car. Approaches of object detection (applied to car, tires, and humans), robust real-time control of robot arms under perception uncertainty, and human-friendly haptic interaction are detailed. The video shows two robots jointly performing the task of mounting a mock-up tire to a real car using the proposed methods, realizing robust performance in an uncertain environment.

While humans can manipulate deformable objects smoothly and naturally, this is still a challenge for autonomous robots due to the complex object dynamics. The presence of rigid environment constraints and altering contact phases between the deformable object, the manipulator, and the environment makes this problem even more challenging. This paper presents a framework for deformable object manipulation that makes use of a single human demonstration of the task. The recorded trajectories are automatically segmented into a sequence of haptic control primitives involving contact with the rigid environment and vision-guided grasp primitives. The recorded motion/force trajectories serve as reference for a compliant control scheme in contact situations. In order to cope with positioning uncertainties a variable admittance control is proposed. The proposed approach is validated in an experimental mounting task for a deformable linear object with multiple re-grasping. The task is demonstrated with a multimodal teleoperation system and transfered to a robotic platform with a pair of seven degrees of freedom manipulators.

The ability to infer intentions and predict actions enables coordinating of one's own actions with those of another human and allows smooth and intuitive interaction. The aim to achieve equally effective human-robot interactions is a crucial aspect of current robotic studies. Thus, we assume that studying human-human interaction provides valuable insights allowing to implement mutual intention recognition and action prediction in robotic systems. A common scenario of interaction, be it in everyday life or in an industrial setting, is that two or more agents share the same workspace and perform tasks without interference. If humans are involved, the robots should act sufficiently predictable to enable the human to attribute goals and predict motion trajectories. In the present work, we first analyzed how well a human recognizes the goal of another person entering the room, and whether this ability is deteriorated by concealing gaze direction of the other person. In a second setup, the same experiment was repeated by replacing the approaching person with a wheeled robot. On average, the distance at which subjects predicted the goal of the approaching agent was approx. 4 m and depended on subject and goal position, but not on the type of agent. However, goal attribution showed a considerable proportion of errors for the robot (19%), much less for a human with hidden gaze direction (6%), and almost none for a human with visible gaze (1%). Thus, our subjects apparently decided on the goal of the approaching agent without taking into account the reliability of directional cues, thus resulting in more errors. In a human-robot setting, such wrong predictions about robotic behavior may easily lead to dangerous situations. For smooth and safe interaction, it is therefore important to ameliorate the predictability of robotic actions.

Natural language interaction between humans and robots is a very challenging topic, especially when it refers to motion descriptions in a certain environment. This problem is particularly relevant during physical human-robot interaction, e.g. in cooperative transportation tasks, where the partners' physical coupling requires an agreement on the way to follow. Understanding in depth the link between sentences, words, environmental properties and motions can deeply enhance the interaction between humans and robots. In this work, we propose a novel approach for learning relations and dependencies between motion, natural language and environmental properties using parameterized left-to-right time-based Hidden Markov Models. A natural language model represents the link between language and motion symbols while the HMMs parameterization corresponds to the explicit influence on motions of both words and environmental features. The proposed PHMM approach parameterizes the output and the transition probabilities using a non-linear dependency estimation. The method is validated by learning and generating navigation primitives in a 2 Degrees-Of-Freedom (DoF) virtual scenario.

Operating in the proximity of humans has been a long-term challenge in robotics research. To achieve this objective, one of the main issues is to ensure safe and comfortable physical human-robot interaction (pHRI). In this paper, we tackle the safety problem at the control level. To ensure operation is within perceived safe zone, we use model predictive control, which finds the optimal control signal online, while imposing predefined safety constraints on the robotic system. The strength of this method lies in allowing the system to perform close to, or at the edge of the constraints' boundaries. In contrast to other works we consider here perceived safety; the constraints for perceived safety are derived in a competitive pHRI experiment. The perceived safety and comfort of the proposed approach is then evaluated with a second, game-like pHRI experiment.

Aiming at the application in physical human-robot interaction, this paper presents a novel adaptive admittance control scheme for robotic manipulators. Special emphasis is drawn on the avoidance of oscillatory behavior in the presence of closed kinematic chains while keeping the rendered impedance low. The approach uses an online fast Fourier transform of the measured manipulator endeffector forces in order to detect oscillations and to adapt the admittance parameters dynamically. As a novel method towards human-centered control design the adaptation strategy is determined in a user study evaluated with a machine-learning algorithm. Experiments conducted with ten human participants show superiority over the non-adaptive admittance control scheme.

Anticipatory behavior based on the human behavior prediction enables the robot to improve the quality of its assistance in physical human-robot interaction (pHRI). However, predictions are partly afflicted with high uncertainties originating from the intrinsic variability in human behavior and the influence of the environment, requiring an attitude negotiation among partners. In this paper, we propose a novel control approach that dynamically adapts the robot's attitude to the disagreement level and the environmental situation in real time facilitating the negotiation between the human and the robot. The approach is based on risk-sensitive optimal feedback control. The adaptive design of the robot's attitude is realized through a dynamical changing risk-sensitivity parameter. The proposed approach is experimentally validated in a cooperative transport scenario in a two-dimensional visuo-haptic virtual environment.

Haptic teleoperation enables the human to perform manipulation tasks in distant, scaled, hazardous, or inaccessible environments. The human closes the control loop sending haptic command signals to and receiving haptic feedback signals from the remote teleoperator. The main research question is how to design the control such that human decision making and action is supported in the best possible way while ensuring robust operation of the system. The human in the loop induces two major challenges for control design: 1) the dynamics of the human operator and the teleoperation system are tightly coupled, i.e., stability of the overall system is affected by the human operator dynamics; and 2) the performance of the teleoperation system is subjectively evaluated by the human, which typically means that standard control performance metrics are not suitable. This paper discusses recent control design successes in the area of haptic teleoperation. In particular, the importance and need of dynamic human haptic closed-loop behavior models and human perception models for the further improvement of haptic teleoperation systems is highlighted and discussed for real-world problem domains.

Audiovisual communications is at the core of multimedia systems that allow users to interact across distances. It is common understanding that both audio and video are required for high-quality interaction. While audiovisual information provides a user with a satisfactory impression of being present in a remote environment, physical interaction and manipulation is not supported. True immersion into a distant environment and efficient distributed collaboration require the ability to physically interact with remote objects and to literally get in touch with other people. Touching and manipulating objects remotely becomes possible if we augment traditional audiovisual communications by the haptic modality. Haptic communications is a relatively young field of research that has the potential to substantially improve human-human and human-machine interaction. In this paper, we discuss the state-of-the-art in haptic communications both from psychophysical and technical points of view. From a human perception point of view, we mainly focus on the multimodal integration of video and haptics and the improved performance that can be achieved when combining them. We also discuss how the human adapts to discrepancies and synchronization errors between different modalities, a research area which is typically referred to as perceptual learning. From a technical perspective, we address perceptual coding of haptic information and the transmission of haptic data streams over resource-constrained and potentially lossy networks in the presence of unpredictable and time-varying communication delays. In this context, we also discuss the need for objective quality metrics for haptic communication. Throughout the paper, we stress the fact that haptic communications is not meant as a replacement of traditional audiovisual communications but rather as an additional dimension for telepresence that will allow us to advance in our quest for truly immersive communication.

This paper is concerned with the asynchronous consensus problem of discrete-time second-order multi-agent system under dynamically changing communication topology, in which the asynchrony means that each agent detects the neighbors' state information to update its state information by its own clock. It is not assumed that the agents' clocks are synchronized. Nor is it assumed that the time sequence over which each agent update its state information is evenly spaced. By using tools from graph theory and nonnegative matrix theory, particularly the product properties of row-stochastic matrices from an infinite set, we finally show that essentially the same result as that for the synchronous discrete-time system holds in the face of asynchronous setting. This generalizes the existing result to a very general case.

High packet rates in telepresence and teleaction systems pose grave challenges to teleoperation over existing communication infrastructure like the Internet. To counter these issues, efficient perceptually motivated packet-rate reduction schemes have been developed. These schemes are conventionally evaluated for perceived quality via subjective user tests. Such tests are time-consuming, expensive and require precise control of experimental conditions. Computer modeling of telepresence sessions can, on the other hand, bring repeatability, ease of observation, definite control over system parameters and task description and fairness of comparison. In this paper, we present first steps towards a methodology and a framework to model and simulate a networked haptic interaction and evaluate it objectively for the quality of experience. Towards this purpose, we model the human control action and haptic perception process in teleoperation. Our results show that simulations of these models for a range of data reduction scheme parameters produce quality estimates whose trend is comparable to carefully performed subjective user tests.

This video presents a robot capable of playing pool on a normal sized pool table using two arms. For successfully completing this task several issues need to be addressed, including the perception of relevant environment information, planning of actions and finally an efficient execution. The video outlines how the robot accurately locates the pool table, the balls on the table and the cue and subsequently plans the next shot. In order to improve the stroke speed, an optimization algorithm for the arm configuration is described. Finally, it is shown how all these modules are integrated to achieve a working two-handed robotic pool play.

In order to speed up image processing in visual servoing, the distributed computational power across networks and appropriate data transmission mechanisms are of particular interest. In this paper, a high sampling rate of visual feedback is achieved by distributed computation on a cloud image processing platform. For target tracking with a networked visual servo control system, a switching control law considering the varying feedback delay caused by image processing and data transmission is applied to improve the control performance. A sending rate scheduling strategy aiming at saving the network load is proposed based on the tracking error. Experiments on a 7 degree-of-freedom (DoF) manipulator are carried out to validate the proposed approach. The proposed approach shows a similar control performance as a system without sending rate scheduling, however, beneficially with largely reduced network load.

This paper explores learning of interaction force skills by human demonstration in dynamic interaction tasks. Skillful force regulation is required in many cases to achieve the goal of a task and at the same time, not to cause undesired stress on the manipulator or the object under manipulation which could result in physical failure. For example, manipulation of compliant objects with varying physical properties or artistic tasks such as engraving require skillful force modulation. Humans gracefully manipulate objects by using their sense of touch and skillfully regulating exerted forces. To learn the demonstrated force for a task by demonstration, an interaction force control policy, in terms of a goal-directed dynamical system, is proposed which stems from the parallel force/position control. The control policy is parameterized and its parameters are learned by Locally Weighted Regression from human demonstrated data to learn a force trajectory. Scaling of learned force is possible by modifying the goal of the system. The proposed method is evaluated in virtual manipulation tasks using a two degrees-of-freedom haptic device.

Physical cooperation with humans greatly enhances the capabilities of robotic systems when leaving standardized industrial settings. Our novel cognition-enabled control framework presented in this paper enables a robotic assistant to enrich its own experience by acquisition of human task knowledge during joint manipulation. Our robot incrementally learns semantic task structures during joint task execution using hierarchically clustered Hidden Markov Models. A semantic labeling of recognized task segments is acquired from the human partner through speech. After a small number of repetitions, the robot uses an anticipated task progress to generate a feed-forward set point for an admittance feedback control scheme. This paper describes the framework and its implementation on a mobile bi-manual platform. The evolution of the robot's task knowledge is presented and discussed. Finally, the cooperation quality is measured in terms of the robot's task contribution.

Synchronization occurs frequently in human behaviour: Everybody has experienced that in a group of people walking pace tends to equalize. The phenomenon of synchrony has been established in the literature in tasks which have little in common with daily life such as pendulum swinging and chair rocking. We extend the knowledge about human movement synchronization by showing that it also occurs during goal-directed actions. In a first experiment, we investigate how synchrony emerges develops over time. In a second experiment, we show that humans also synchronize their actions with a robot. Results are interpreted in the light of joint action theory. Possible implications and improvements for human-robot interaction are discussed.

In this paper, a networked visual servo control system with distributed computation is proposed to overcome the low sampling rate problem in vision-based control systems. A real-time image data transmission protocol based on Realtime Transport Protocol (RTP) is developed. The captured images are sent to different processing nodes connected over a communication network and processed in parallel. Thus, a high sampling rate of the visual feedback is achieved under a cloud image processing architecture. The varying image processing delay caused by the varying number of extracted features and the random transmission delay are modeled as a random process with Bernoulli distribution. By using the input-delay approach, the resulted networked visual servo control system is reformulated into a stochastic continuous-time system with time-varying delay. Experiments on two 1-DoF linear motor modules are carried out to validate the proposed approach. A visual servo control system without parallel distributed computation is implemented for comparison. The experimental results demonstrate significant performance improvement by the proposed approach.

The collaborative research center SFB453 (www.sfb453.de) aims to realize high-fidelity telepresence and teleaction systems. Telepresence and teleaction systems extend the human workspace to remote locations in order to overcome barriers like distance, scaling, danger or the human skin. Using a human-system interface the human operator controls a remotely located teleoperator. Multi-modal feedback in form of visual, auditory, and haptic data is used to increase the feeling of telepresence. Different application areas including minimally invasive surgery, on-orbit servicing, microassembly as well as tele-manufacturing and tele-maintenance are targeted.

This work aims at the development of a versatile control strategy for operating unknown mechanically constrained devices such as drawers or doors. Few assumptions on the device's shape as well as the utilized hardware are required. Our approach is based on an on-line estimation of the constraint manifold which serves as a reference input for an admittance-type controller providing the compliance required. The direction estimation is obtained from the velocity signal in task space. An on-line adaptation of the admittance controller according to the estimated moving direction reduces contact forces. The functionality of the control strategy is demonstrated on a mobile manipulator in a kitchen environment.

In this paper, a novel switching controller is proposed for a networked visual servo control system with varying feedback delay due to image processing and data transmission. The varying image processing delay caused by the varying number of extracted features for pose estimation due to different view angles, illumination conditions and noise, is modeled by its occurrence probability. The time delay due to transmission over the communication network is also modeled as random process. By using a sampled-data system approach and an input-delay approach, the linearized visual servo control system is reformulated into a stochastic continuous-time system with time-varying delay. A novel stability condition and associated switching controller are derived based on the occurrence probabilities of delays. Experiments on a 1-DoF linear module equipped with a camera are conducted to validate the proposed approach. A non-switching controller approach is implemented for comparison. The experimental results demonstrate significant performance improvement of the proposed control approach.

The overall performance of a robotic system is commonly expressed by a single scenario-specific metric which is supposed to be optimized. However, the metric describing the performance of a single subtask within a scenario may be different. Nevertheless, the scenario performance is most likely dependent on the subtask performances but a mutual transformation is not straightforward in general, especially in complex robotic systems. This leads to what we call the common pricing problem, i.e. the problem to determine the functional relationship among a set of different performance criteria and then account for this relationship in the various optimizations throughout all system layers. In this paper we present an approach to first learn a probabilistic model of the metric interdependencies, and thereafter utilize this model for performance estimation and optimal task parameterization during planning and execution respectively. The proposed method is validated in a simulation.

The stability and performance of visual servo control systems strongly depend on the delays caused by image processing. In order to accelerate the visual feedback, the distributed computational power across networks and appropriate data transmission mechanism are of particular interest. In this paper, a novel distributed computation with data scheduling is proposed for networked visual servo control systems (NVSCSs) aiming at improving the control performance. A realtime transport protocol is developed for image data transmission. For a NVSCS which is modeled as a continuous-time system with computation, transmission and holding delays, a switching control law is applied. A probabilistic sampling scheduler is derived such that the control performance and the network load caused by image data transmission are balanced. Experiments on two 1-DoF linear modules equipped with a camera are conducted to validate the proposed approach. A visual servo system without data scheduling is implemented for comparison. The experimental results demonstrate a comparable control performance of the proposed approach with an advantage of reduced network load.

Physical cooperation with humans greatly enhances the capabilities of robotic systems when leaving standardized industrial settings. In particular, manipulation of bulky objects in narrow environments requires cooperating partners. Actuation redundancies arising in joint manipulation impose the question of load sharing among the interacting partners. In this paper, effort sharing policies are systematically derived from the geometric and dynamic task properties. Three policies are intuitively identified, resulting in unilateral and balanced effort distributions. These policies are evaluated within a novel hierarchical motion generation and control framework. The synthesized system is successfully validated in a three-degrees-of-freedom planar tracking experiment. This evaluation shows an interdependency of the load sharing strategy and the resulting task performance.

Telesurgery systems integrate multimodal communication and robotic technologies to enable surgical procedures to be performed from remote locations. They allow human surgeons to intuitively control laparoscopic instruments and to navigate within the human body. In this paper, we present selected topics on multimodal interaction in the context of telesurgery applications. These are results from the collaborative research project SFB 453 on ldquoHigh-Fidelity Telepresence and Teleactionrdquo which is funded by the German Research Foundation in the larger Munich area. The focus in this paper is on multimodal information processing and communication including simulation of surgical targets in the human body. Furthermore, we present an overview of our advanced multimodal telesurgery demonstrators that provide a comprehensive platform for our collaborative telepresence research.

In multimodal tele-cooperation as considered in this paper two humans in distant locations jointly perform a task requiring multimodal including haptic feedback. One human operator teleoperates a remotely placed humanoid robot which is collocated with the human cooperator. Time delay in the communication channel as destabilizing factor is one of the multiple challenges associated with such a tele-cooperation setup. In this paper we employ a control architecture with force-position exchange accounting for the admittance type of the haptic input device and the telerobot, which both are position-based admittance controlled. Llewellynpsilas stability criteria are employed for the parameter tuning of the virtual impedances in the presence of time delay. The control strategy is successfully validated in an intercontinental tele-cooperation experiment with the humanoid telerobot HRP-2 located in Japan/Tsukuba and a multimodal human-system-interface located in Germany/Munich, see also the corresponding video submission. The proposed setup gives rise to a large number of exciting new research questions to be addressed in the future.

The video shows an intercontinental cooperative telemanipulation task, whereby the operator site is located in Munich, Germany and the teleoperator site in Tsukuba, Japan. The human operator controls a remotely located teleoperator, which performs a task in the remote environment. Hereby the human operator is assisted by another person located at the remote site. The task consists in jointly grasping an object, moving it to a new position and finally releasing it, see Fig. 1.

Telepresence systems are often deployed in scenarios where communication bandwidth is limited. Consequently, data exchanged between operator and teleoperator has to be reduced. In case of haptic telepresence, data reduction has an influence on the stability of the overall system. This paper provides a step towards a systematic framework for communication data bandwidth reduction in haptic telepresence systems discussing stability for a class of lossy data reduction (LDR) algorithms. Simulation and experimental results validate the efficacy

One of the key challenges in telepresence and teleaction systems is the fact that a global control loop is closed over a communication network. The transmission delay of haptic information is extremely critical. Therefore, new data samples from the haptic sensors are typically immediately forwarded to the receiver which leads to a large number of packets being generated when using the Internet as the communication infrastructure. We present a novel approach to reduce the number of packets and, therefore, the amount data communicated in a telepresence and teleaction system. Our method uses a passive deadband transmission approach which only delivers data packets over the network when the sampled sensor data changes more than a given threshold value. The threshold value is determined by psychophysical experiments. This approach leads to a considerable reduction (up to 90%) of packet rate and data rate without sacrificing the fidelity and immersiveness of the system.