There seems to be a solid object inside.""What other toppings do you usually like on your pizza?" Figure 1: We imbued a robot with curious behaviors.The figure shows two examples.Left: The robot shakes a container to check whether there is an object inside.Right: The robot asks for the person's preferences.

Imitation Learning offers a promising approach to learn directly from data without requiring explicit models, simulations, or detailed task definitions. During inference, actions are sampled from the learned distribution and executed on the robot. However, sampled actions may fail for various reasons, and simply repeating the sampling step until a successful action is obtained can be inefficient. In this work, we propose an enhanced sampling strategy that refines the sampling distribution to avoid previously unsuccessful actions. We demonstrate that by solely utilizing data from successful demonstrations, our method can infer recovery actions without the need for additional exploratory behavior or a high-level controller. Furthermore, we leverage the concept of diffusion model decomposition to break down the primary problem—which may require long-horizon history to manage failures—into multiple smaller, more manageable sub-problems in learning, data collection, and inference, thereby enabling the system to adapt to variable failure counts. Our approach yields a low-level controller that dynamically adjusts its sampling space to improve efficiency when prior samples fall short. We validate our method across several tasks, including door opening with unknown directions, object manipulation, and button-searching scenarios, demonstrating that our approach outperforms traditional baselines. Supplementary materials for this paper are available on our website: https://hri-eu.github.io/ccdp/.

Robot-assisted dressing is a popular but challenging topic in the field of robotic manipulation, offering significant potential to improve the quality of life for individuals with mobility limitations. Currently, the majority of research on robot-assisted dressing focuses on how to put on loose-fitting clothing, with little attention paid to tight garments. For the former, since the armscye is larger, a single robotic arm can usually complete the dressing task successfully. However, for the latter, dressing with a single robotic arm often fails due to the narrower armscye and the property of diminishing rigidity in the armscye, which eventually causes the armscye to get stuck. This paper proposes a bimanual dressing strategy suitable for dressing tight-fitting clothing. To facilitate the encoding of dressing trajectories that adapt to different human arm postures, a spherical coordinate system for dressing is established. We uses the azimuthal angle of the spherical coordinate system as a task-relevant feature for bimanual manipulation. Based on this new coordinate, we employ Gaussian Mixture Model (GMM) and Gaussian Mixture Regression (GMR) for imitation learning of bimanual dressing trajectories, generating dressing strategies that adapt to different human arm postures. The effectiveness of the proposed method is validated through various experiments.

The gaze of a person tends to reflect their interest. This work explores what happens when this statement is taken literally and applied to robots. Here we present a robot system that employs a moving robot head with a screen-based eye model that can direct the robot’s gaze to points in physical space and present a reflection-like mirror image of the attended region on top of each eye. We conducted a user study with 33 participants, who were asked to instruct the robot to perform pick-and-place tasks, monitor the robot’s task execution, and interrupt it in case of erroneous actions. Despite a deliberate lack of instructions about the role of the eyes and a very brief system exposure, participants felt more aware about the robot’s information processing, detected erroneous actions earlier, and rated the user experience higher when eye-based mirroring was enabled compared to non-reflective eyes. These results suggest a beneficial and intuitive utilization of the introduced method in cooperative human-robot interaction.

In the pursuit of fully autonomous robotic systems capable of taking over tasks traditionally performed by humans, the complexity of open-world environments poses a considerable challenge. Addressing this imperative, this study contributes to the field of Large Language Models (LLMs) applied to task and motion planning for robots. We propose a system architecture that orchestrates a seamless interplay between multiple cognitive levels, encompassing reasoning, planning, and motion generation. At its core lies a novel replanning strategy that handles physically grounded, logical, and semantic errors in the generated plans. We demonstrate the efficacy of the proposed feedback architecture, particularly its impact on executability, correctness, and time complexity via empirical evaluation in the context of a simulation and two intricate real-world scenarios: blocks world, barman and pizza preparation.

Learning complex trajectories from demonstrations in robotic tasks has been effectively addressed through the utilization of Dynamical Systems (DS). State-of-the-art DS learning methods ensure stability of the generated trajectories; however, they have three shortcomings: a) the DS is assumed to have a single attractor, which limits the diversity of tasks it can achieve, b) state derivative information is assumed to be available in the learning process and c) the state of the DS is assumed to be measurable at inference time. We propose a class of provably stable latent DS with possibly multiple attractors, that inherit the training methods of Neural Ordinary Differential Equations, thus, dropping the dependency on state derivative information. A diffeomorphic mapping for the output and a loss that captures time-invariant trajectory similarity are proposed. We validate the efficacy of our approach through experiments conducted on a public dataset of handwritten shapes and within a simulated object manipulation task.

How can a robot provide unobtrusive physical support within a group of humans? We present Attentive Support, a novel interaction concept for robots to support a group of humans. It combines scene perception, dialogue acquisition, situation understanding, and behavior generation with the common-sense reasoning capabilities of Large Language Models (LLMs). In addition to following user instructions, Attentive Support is capable of deciding when and how to support the humans, and when to remain silent to not disturb the group. With a diverse set of scenarios, we show and evaluate the robot’s attentive behavior, which supports and helps the humans when required, while not disturbing if no help is needed.

This paper investigates one of the most challenging tasks in dynamic manipulation-catching large-momentum moving objects. Beyond the realm of quasi-static manipulation, dealing with highly dynamic objects can significantly improve the robot's capability of interacting with its surrounding environment. Yet, the inevitable motion mismatch between the fast moving object and the approaching robot will result in large impulsive forces, which lead to the unstable contacts and irreversible damage to both the object and the robot. To address the above problems, we propose an online optimization framework to: 1) estimate and predict the linear and angular motion of the object; 2) search and select the optimal contact locations across every surface of the object to mitigate impact through sequential quadratic programming (SQP); 3) simultaneously optimize the end-effector motion, stiffness, and contact force for both robots using multi-mode trajectory optimization (MMTO); and 4) realise the impact-aware catching motion on the compliant robotic system based on indirect force controller. We validate the impulse distribution, contact selection, and impactaware MMTO algorithms in simulation and demonstrate the benefits of the proposed framework in real-world experiments including catching large-momentum moving objects with welldefined motion, constrained motion and free-flying motion.

Developing physically assistive robots capable of dressing assistance has the potential to significantly improve the lives of the elderly and disabled population. However, most robotics dressing strategies considered a single robot only, which greatly limited the performance of the dressing assistance. In fact, healthcare professionals perform the task bimanually. Inspired by them, we propose a bimanual cooperative scheme for robotic dressing assistance. In the scheme, an interactive robot joins hands with the human thus supporting/guiding the human in the dressing process, while the dressing robot performs the dressing task. We identify a key feature: elbow angle that affects the dressing action and propose an optimal strategy for the interactive robot using the feature. A dressing coordinate based on the posture of the arm is defined to better encode the dressing policy. We validate the interactive dressing scheme with extensive experiments and also an ablation study. The experiment video is available onhttps://sites.google.com/view/bimanualassitdressing/home

Physically assistive robots and exoskeletons have great potential to help humans with a wide variety of collaborative tasks. However, a challenging aspect of the control of such devices is to accurately model or predict human behaviour, which can be highly individual and personalised. In this work, we implement a framework for learning subject-specific models of underlying human motion strategies using inverse musculoskeletal optimal control. We apply this framework to a specific motion task: the sit-to-stand transition. By collecting sit-to-stand data from 4 subjects with and without perturbations, we show that humans modulate their sit-to-stand strategy in the presence of instability, and learn the corresponding models of these strategies. In the future, the personalised motion strategies resulting from this framework could be used to inform the design of real-time assistance strategies for human-robot collaboration problems.

Symbolic planning is a powerful technique to solve complex tasks that require long sequences of actions and can equip an intelligent agent with complex behavior. The downside of this approach is the necessity for suitable symbolic representations describing the state of the environment as well as the actions that can change it. Traditionally such representations are carefully hand-designed by experts for distinct problem domains, which limits their transferability to different problems and environment complexities. In this paper, we propose a novel concept to generalize symbolic actions using a given entity hierarchy and observed similar behavior. In a simulated grid-based kitchen environment, we show that type-generalized actions can be learned from few observations and generalize to novel situations. Incorporating an additional on-the-fly generalization mechanism during planning, unseen task combinations, involving longer sequences, novel entities and unexpected environment behavior, can be solved.

Learning robot control policies from physics simulations is of great interest to the robotics community as it may render the learning process faster, cheaper, and safer by alleviating the need for expensive real-world experiments. However, the direct transfer of learned behavior from simulation to reality is a major challenge. Optimizing a policy on a slightly faulty simulator can easily lead to the maximization of the 'Simulation Optimization Bias' (SOB). In this case, the optimizer exploits modeling errors of the simulator such that the resulting behavior can potentially damage the robot. We tackle this challenge by applying domain randomization, i.e., randomizing the parameters of the physics simulations during learning. We propose an algorithm called Simulation-based Policy Optimization with Transferability Assessment (SPOTA) which uses an estimator of the SOB to formulate a stopping criterion for training. The introduced estimator quantifies the over-fitting to the set of domains experienced while training. Our experimental results on two different second order nonlinear systems show that the new simulation-based policy search algorithm is able to learn a control policy exclusively from a randomized simulator, which can be applied directly to real systems without any additional training.

This paper presents a method to incorporate ergonomics into the optimization of action sequences for bi-manual human-robot cooperation tasks with continuous physical interaction. Our first contribution is a novel computational model of the human that allows prediction of an ergonomics assessment corresponding to each step in a task. The model is learned from human motion capture data in order to predict the human pose as realistically as possible. The second contribution is a combination of this prediction model with an informed graph search algorithm, which allows computation of human-robot cooperative plans with improved ergonomics according to the incorporated method for ergonomic assessment. The concepts have been evaluated in simulation and in a small user study in which the subjects manipulate a large object with a 32 DoF bimanual mobile robot as partner. For all subjects, the ergonomic-enhanced planner shows their reduced ergonomic cost compared to a baseline planner.

The transition from free motion to contact is a challenging problem in robotics, in part due to its hybrid nature. Additionally, disregarding the effects of impacts at the motion planning level often results in intractable impulsive contact forces. In this paper, we introduce an impact-aware multi-mode trajectory optimization (TO) method that combines hybrid dynamics and hybrid control in a coherent fashion. A key concept is the incorporation of an explicit contact force transmission model in the TO method. This allows the simultaneous optimization of the contact forces, contact timings, continuous motion trajectories and compliance, while satisfying task constraints. We compare our method against standard compliance control and an impact-agnostic TO method in physical simulations. Further, we experimentally validate the proposed method with a robot manipulator on the task of halting a large-momentum object.

Effective collaboration is based on online adaptation of one's own actions to the actions of their partner. This article provides a principled formalism to address online adaptation in joint planning problems such as Dyadic collaborative Manipulation (DcM) scenarios. We propose an efficient bilevel formulation that combines graph search methods with trajectory optimization, enabling robotic agents to adapt their policy on-the-fly in accordance to changes of the dyadic task. This method is the first to empower agents with the ability to plan online in hybrid spaces; optimizing over discrete contact locations, contact sequence patterns, continuous trajectories, and force profiles for co-manipulation tasks. This is particularly important in large object co-manipulation that requires changes of grasp-holds and plan adaptation. We demonstrate in simulation and with robot experiments the efficacy of the bilevel optimization by investigating the effect of robot policy changes in response to real-time alterations of the dyadic goals, eminent grasp switches, as well as optimal dyadic interactions to realize the joint task.

The work undertaken in this paper was done in the Centre of Excellence for Global Systems Science (CoeGSS), an interdisciplinary project, funded by the European Commission.

Monitoring of High Performance Computing (HPC) platforms is critical to successful operations, can provide insights into performance-impacting conditions, and can inform methodologies for improving science throughput. However, monitoring systems are not generally considered core capabilities in system requirements specifications nor in vendor development strategies. In this paper we present work performed at a number of large-scale HPC sites towards developing monitoring capabilities that fill current gaps in ease of problem identification and root cause discovery. We also present our collective views, based on the experiences presented, on needs and requirements for enabling development by vendors or users of effective sharable end-to-end monitoring capabilities.

This paper presents a system for cooperatively manipulating large objects between a human and a robot. This physical interaction system is designed to handle, transport, or manipulate large objects of different shapes in cooperation with a human. Unique points are the bi-manual physical cooperation, the sequential characteristic of the cooperation including contact changes, and a novel architecture combining force interaction cues, interactive search-based planning, and online trajectory and motion generation. The resulting system implements a mixed initiative collaboration strategy, deferring to the human when his intentions are unclear, and driving the task once understood. This results in an easy and intuitive human-robot interaction. It is evaluated in simulations and on a bi-manual mobile robot with 32 degrees of freedom.

Learning sequential force interaction tasks from kinesthetic demonstrations is a promising approach to transfer human manipulation abilities to a robot. In this paper we propose a novel concept to decompose such demonstrations into a set of Movement Primitives (MPs). The decomposition is based on a probability distribution we call Directional Normal Distribution (DND). To capture the sequential properties of the manipulation task, we model the demonstrations with a Hidden Markov Model (HMM). Here, we employ mixtures of DNDs as the HMM's output emissions. The combination of HMMs and mixtures of DNDs allows to infer the MP's composition, i.e., its coordinate frames, control variables and target coordinates from the demonstration data. In addition, it permits to determine an appropriate number of MPs that explains the demonstrations best. We evaluate the approach on kinesthetic demonstrations of a light bulb unscrewing task. Decomposing the task leads to intuitive and meaningful MPs that reflect the natural structure of the task.

With the increasing popularity of the cloud computing model and rapid proliferation of cloud infrastructures there are increasing concerns about energy consumption and consequent impact of cloud computing as a contributor to global CO2emissions. To date, little is known about how to incorporate energy consumption and CO2concerns into cloud application development and deployment decision models. In this respect, this paper describes an eco-aware approach that relies on the definition, monitoring and utilization of energy and CO2metrics combined with the use of innovative application scheduling and runtime adaptation techniques to optimize energy consumption and CO2footprint of cloud applications as well as the underlying infrastructure. The eco-aware approach involves measuring or quantifying the energy consumption and CO2at different levels of cloud computing, using that information to create scheduling and adaptation techniques that contribute towards reducing the energy consumption and CO2emissions, and finally testing and validating the developed solutions in a multi-site cloud environment with the help of challenging case study applications. The experimental and validation results show the potential of the eco-aware approach to significantly reduce the CO2footprint and consequent environmental impact of cloud applications.

Kinesthetic teaching is a promising approach to acquire robot skills in an intuitive way. This paper focuses on learning skills that do not solely rely on kinematics but also need to take into account interaction forces. We present three novel concepts towards learning such force interaction skills. Firstly, we determine segments from a small number of continuous kinesthetic demonstrations using contact information. Secondly, we associate each segment with a movement primitive, and determine its composition, i.e., the control variables and reference frames that allow to reproduce the demonstrated task. Lastly, we propose a concept to determine the transitions between the primitives during reproduction. The proposed methods are evaluated on a box pulling and flipping task, and show very good generalization abilities for objects with different geometries, and situations with different object arrangements.

This paper presents an effective method for planning and optimizing robot motions in joint space using via-points. The via-point formulation allows for a sparse movement representation which is inherently smooth according to a minimum jerk criterion. In this research we focus on two aspects. First, we present an initialization method to find a feasible number of via-points and their distribution in time. This initialization takes the difficulty of the planning problem into account and finds a tradeoff between representation complexity (number of via-points) and optimization time. Second, we propose a two-step optimization approach to improve the via-point parameters to satisfy constraints over the whole motion duration. We show that our methods generate feasible and smooth robot motions with a high success rate for tasks of varying complexity. The methods are evaluated in simulation using four different scenarios with two different robots, a 7-DOF robot arm and a 16-DOF humanoid upper body. Additionally, the experiments with the robot arm are shown in a real-world experiment.

Classical approaches towards learning coordinated movement tasks often represent a movement in a sequential and exclusive fashion. Introducing concurrency allows to decompose such tasks into a number of separate sequences, for instance for two different end-effectors. While this results in a compact and generic representation of the individual movement primitives (MPs), it is a hard problem to learn their temporal and causal organization. This paper presents a concept for learning movement tasks that require the coordination of several controlled effectors of a robot. We firstly introduce a concept to learn and estimate the progress of individual MPs from a low number of demonstrations. Secondly, we propose a representation of the task that incorporates several concurrent sequences of MPs. Combining these two elements allows to learn and reproduce coordinated bi-manual movement tasks robustly. The synchronization of the concurrent MPs is achieved implicitly using the progress prediction. The approach is evaluated in two simulation studies with a 25 degrees of freedom two-arm robot performing a pick-and-place task.

We present an approach for learning sequential robot skills through kinesthetic teaching. The demonstrations are represented by a sequence graph. Finding the transitions between consecutive basic movements is treated as classification problem where both Support Vector Machines and Gaussian Mixture Models are evaluated as classifiers. We show how the observed primitive order of all demonstrations can help to improve the movement reproduction by restricting the classification outcome to the currently executed primitive and its possible successors in the graph. The approach is validated with an experiment in which a 7-DOF Barrett WAM robot learns to unscrew a light bulb.

This paper introduces a motion generation framework that integrates a hierarchical movement primitive (MP) layer with optimal control in form of receding horizon optimization. In order to benefit from fast reactions on the MP-layer, the optimal control layer can be overridden in risky situations to generate quick, though non-optimal solutions. By this, the system fulfills four desirable properties. It continuously adapts the robot's motion without noticeable delay (1) by optimizing for collision and joint limit avoidance based on a future time horizon instead of the current state only (2). It accounts for the full robot motion that may result from multiple active MPs at the same time (3) and despite a possibly slow optimization still provides the robustness and quick reaction capabilities of MPs (4). The framework has been validated in an experiment in which a humanoid robot performed a task, optimized wrt. collisions and joint limit avoidance, but still could react within 50 ms after detection of a potential risk.

BonFIRE is a multi-site test bed that supports testing of Cloud-based and distributed applications. BonFIRE breaks the mould of commercial Cloud offerings by providing unique functionality in terms of observability, control, advanced Cloud features and ease of use for experimentation. A number of successful use cases have been executed on BonFIRE, involving industrial and academic users and delivering impact in diverse areas, such as media, e-health, environment and manufacturing. The BonFIRE user-base is expanding through its free, Open Access scheme, daily carrying out important research, while the consortium is working to sustain the facility beyond 2014.

The perception of the surrounding world depends on noisy sensors which introduce uncertainty. When we develop algorithms for grasping with robotic hands it is not enough to assume the best estimate of the environment - if there is a measure of uncertainty we need to account for it. This paper presents a control law which augments a grasp controller with the ability to prefer known or unseen regions of an object; this leads to the introduction of two motion primitives: an explorative and exploitative grasp. We integrate this control law in a framework for iterative grasping and implement a tactile exploration scenario. The experimental results confirm that using the notion of uncertainty in the control loop yields better models and does it faster than an uninformed controller.

The development and maintenance of cloud sites are often characterized by energy waste and high CO2 emissions. Energy efficiency and the decrease of the CO2 emissions in cloud-based systems can be only obtained by adopting suitable actions and techniques (e.g., utilization of green energy sources, reduction of the number of physical and virtual machines, usage of the greener machines). In order to evaluate the suitability of these different actions, it is necessary to define a measure for greenness of the whole system. For this reason, this paper defines a set of metrics to assess the greenness of a cloud infrastructure. In order to provide a detailed view of the behaviour of the system and to facilitate the identification of the causes of the energy waste, metrics have been defined at different layers of the system (i.e., application, virtualization, infrastructure layers). The monitoring infrastructure that is necessary to retrieve all the data required for the assessment of the identified set of metrics is also described.

This paper addresses the question of how to create adaptive and smooth sequences of actions and how to decide among skill options in a continuous manner without the necessity of recurrent planning. Motion generation is based on serial and parallel blending of movement primitives (MP). MPs are modeled as dynamical systems on task coordinates with attractor behavior and augmented with additional signals to ease their coordination. Sequences and transitions between skills are realized in a unified way as bifurcating dynamical systems based on continuous-time recurrent neural networks. The neural output is used as activation signal for MPs. Besides continuous feedback from the controlled MPs, the neural dynamics is influenced by a cost term from a future prediction to allow the inhibition of an action flow that is expected to fail. First results are shown in a physical simulation environment on a high-DoF robotic hand-arm system. The system is capable of creating smooth transients of MPs. Robustness to disturbances can be observed as local adaptations of individual low-level MPs, flexible sequencing of MPs, and global error recovery by changing the whole strategy of how to perform a movement skill.

The choice of an adequate object shape representation is critical for efficient grasping and robot manipulation. A good representation has to account for two requirements: it should allow uncertain sensory fusion in a probabilistic way and it should serve as a basis for efficient grasp and motion generation. We consider Gaussian process implicit surface potentials as object shape representations. Sensory observations condition the Gaussian process such that its posterior mean defines an implicit surface which becomes an estimate of the object shape. Uncertain visual, haptic and laser data can equally be fused in the same Gaussian process shape estimate. The resulting implicit surface potential can then be used directly as a basis for a reach and grasp controller, serving as an attractor for the grasp end-effectors and steering the orientation of contact points. Our proposed controller results in a smooth reach and grasp trajectory without strict separation of phases. We validate the shape estimation using Gaussian processes in a simulation on randomly sampled shapes and the grasp controller on a real robot with 7DoF arm and 7DoF hand.

Humanoid robots are intended to act and interact in dynamically changing environments in the presence of humans. Current robotic systems are usually able to move in dynamically changing environments because of an inbuilt depth and obstacle sensing. However, for acting in their environment the internal representation of such systems is usually constructed by hand and known in advance. In contrast, this paper presents a system that dynamically constructs its internal scene representation using a model-based vision approach. This enables our system to approach and grasp objects in an previously unknown scene. We combine standard stereo with model-based image fitting techniques for a real-time estimation of the position and orientation of objects. The model-based image processing allows for an easy transfer to the internal, dynamic scene representation. For movement generation we use a task-level whole-body control approach that is coupled with a movement optimization scheme. Furthermore, we present a novel method that constrains the robot to keep certain objects in the FOV while moving. We demonstrate the successful interplay between model-based vision, dynamic scene representation, and movement generation by means of some interactive reaching and grasping tasks.

In this paper we present an approach to reduce the effort for planning robot motions by shifting the planning problem to a high-level representation. We combine classical sampling-based random tree planning with a reactive controller connecting sampling points with nontrivial trajectories, utilizing redundant DOFs to locally avoid obstacles. While the reactive planner operates locally on a short time scale, the complementary sampling-based method is able to find globally feasible solutions due to its larger preview horizon. Additionally, planning is done in a low-dimensional task space instead of the high-dimensional joint space. Comparing the average planning time and number of tree extensions for several scenarios and planning methods, we demonstrate that this hybrid planning approach is capable of solving a large fraction of planning queries while saving considerable planning time.

This paper presents an architecture for learning and reproducing movements with a robot in interaction with a human teacher. We focus on the movement representation and propose three enhancements to increase generalization capabilities: Firstly, we introduce a flexible task-level movement representation that is based on neuropsychological findings. Movement is represented in task-oriented frames of reference, and generalizes to a variety of different situations. Secondly, we propose a mechanism to decouple the task descriptors from the perceived objects in the robot's environment. This allows to formulate a set of generic controllers, and to interactively create associations with perceived objects. Thirdly, we introduce a method to dynamically modify the system's body schema to account for structural changes such as having grasped a tool. The changes are consistently treated in the kinematics computations. This permits to generalize movements to be carried out in different ways, for instance with different hands or bi-manually. A set of experiments in an interactive imitation learning situation underline the capabilities of the proposed concepts.

In this paper we present a new robot control and learning framework. By integrating previously presented as well as new methods, the robot is able to learn an invariant and generic movement representation from a human tutor. We argue that in order to apply such generic representations to new situations and thus create a flexible system, the use of interaction is beneficial. The interaction is based on a kinematically controlled model of a human tutor, which is used as a model-based filter and also for recognizing postures that influence the interaction. In addition, a new movement segmentation scheme is presented that is based on correlating movements by the tutor's hand with the salient objects in the scene. The focus of this paper is on the interactive learning aspects of the system and particular emphasis is given to an experiment in which the humanoid robot ASIMO learns from a human tutor. The system includes extensive generalization capabilities that result from an online adaption of the robot's body schema and the exploitation of inter-trial variance from multiple demonstrations. This enables the robot to reproduce the movement in new situations. For example, a stacking task that the tutor performed one-handed can be executed bi-manually by the robot.

Many everyday human skills can be framed in terms of performing some task subject to constraints imposed by the environment. Constraints are usually unobservable and frequently change between contexts. In this paper, we present a novel approach for learning (unconstrained) control policies from movement data, where observations come from movements under different constraints. As a key ingredient, we introduce a small but highly effective modification to the standard risk functional, allowing us to make a meaningful comparison between the estimated policy and constrained observations. We demonstrate our approach on systems of varying complexity, including kinematic data from the ASIMO humanoid robot with 27 degrees of freedom.

Recent advances in the field of humanoid robotics increase the complexity of the tasks that such robots can perform. This makes it increasingly difficult and inconvenient to program these tasks manually. Furthermore, humanoid robots, in contrast to industrial robots, should in the distant future behave within a social environment. Therefore, it must be possible to extend the robot's abilities in an easy and natural way. To address these requirements, this work investigates the topic of imitation learning of motor skills. The focus lies on providing a humanoid robot with the ability to learn new bi-manual tasks through the observation of object trajectories. For this, an imitation learning framework is presented, which allows the robot to learn the important elements of an observed movement task by application of probabilistic encoding with Gaussian Mixture Models. The learned information is used to initialize an attractor-based movement generation algorithm that optimizes the reproduced movement towards the fulfillment of additional criteria, such as collision avoidance. Experiments performed with the humanoid robot ASIMO show that the proposed system is suitable for transferring information from a human demonstrator to the robot. These results provide a good starting point for more complex and interactive learning tasks.

Many everyday human skills can be framed in terms of performing some task subject to constraints imposed by the environment. Constraints are usually unobservable and frequently change between contexts. In this paper, we present a novel approach for learning (unconstrained) control policies from movement data, where observations are recorded under different constraint settings. Our approach seamlessly integrates unconstrained and constrained observations by performing hybrid optimisation of two risk functionals. The first is a novel risk functional that makes a meaningful comparison between the estimated policy and constrained observations. The second is the standard risk, used to reduce the expected error under impoverished sets of constraints. We demonstrate our approach on systems of varying complexity, and illustrate its utility for transfer learning of a car washing task from human motion capture data.

Previous work [1] shows that the movement representation in task spaces offers many advantages for learning object-related and goal-directed movement tasks through imitation. It allows to reduce the dimensionality of the data that is learned and simplifies the correspondence problem that results from different kinematic structures of teacher and robot. Further, the task space representation provides a first generalization, for example wrt. differing absolute positions, if bi-manual movements are represented in relation to each other. Although task spaces are widely used, even if they are not mentioned explicitly, they are mostly defined a priori. This work is a step towards an automatic selection of task spaces. Observed movements are mapped into a pool of possibly even conflicting task spaces and we present methods that analyze this task space pool in order to acquire task space descriptors that match the observation best. As statistical measures cannot explain importance for all kinds of movements, the presented selection scheme incorporates additional criteria such as an attention-based measure. Further, we introduce methods that make a significant step from purely statistically-driven task space selection towards model-based movement analysis using a simulation of a complex human model. Effort and discomfort of the human teacher is being analyzed and used as a hint for important task elements. All methods are validated with real-world data, gathered using color tracking with a stereo vision system and a VICON motion capturing system.

Based on inspirations from infant development we present a system which learns associations between acoustic labels and visual representations in interaction with its tutor. The system is integrated with a humanoid robot. Except for a few trigger phrases to start learning all acoustical representations are learned online and in interaction. Similar, for the visual domain the clusters are not predefined and fully learned online. In contrast to other interactive systems the interaction with the acoustic environment is solely based on the two microphones mounted on the robots head. In this paper we give an overview on all key elements of the system and focus on the challenges arising from the headset-free learning of speech labels. In particular we present a mechanism for auditory attention integrating bottom-up and top-down information for the segmentation of the acoustic stream. The performance of the system is evaluated based on offline tests of individual parts of the system and an analysis of the online behavior.

Many advanced motion control strategies developed in robotics use captured human motion data as valuable source of examples to simplify the process of programming or learning complex robot motions. Direct and online control of robots from observed human motion has several inherent challenges. The most important may be the representation of the large number of mechanical degrees of freedom involved in the execution of movement tasks. Attempting to map all such degrees of freedom from a human to a humanoid is a formidable task from an instrumentation and sensing point of view. More importantly, such an approach is incompatible with mechanisms in the central nervous system which are believed to organize or simplify the control of these degrees of freedom during motion execution and motor learning phase. Rather than specifying the desired motion of every degree of freedom for the purpose of motion control, it is important to describe motion by low dimensional motion primitives that are defined in Cartesian (or task) space. In this paper, we formulate the human to humanoid retargeting problem as a task space control problem. The control objective is to track desired task descriptors while satisfying constraints such as joint limits, velocity limits, collision avoidance, and balance. The retargeting algorithm generates the joint space trajectories that are commanded to the robot. We present experimental and simulation results of the retargeting control algorithm on the Honda humanoid robot ASIMO.

Transferring motion from a human demonstrator to a humanoid robot is an important step toward developing robots that are easily programmable and that can replicate or learn from observed human motion. The so called motion retargeting problem has been well studied and several off-line solutions exist based on optimization approaches that rely on pre-recorded human motion data collected from a marker-based motion capture system. From the perspective of human robot interaction, there is a growing interest in online and marker-less motion transfer. Such requirements have placed stringent demands on retargeting algorithms and limited the potential use of off-line and pre-recorded methods. To address these limitations, we present an online task space control theoretic retargeting formulation to generate robot joint motions that adhere to the robot’s joint limit constraints, self-collision constraints, and balance constraints. The inputs to the proposed method include low dimensional normalized human motion descriptors, detected and tracked using a vision based feature detection and tracking algorithm. The proposed vision algorithm does not rely on markers placed on anatomical landmarks, nor does it require special instrumentation or calibration. The current implementation requires a depth image sequence, which is collected from a single time of flight imaging device. We present online experimental results of the entire pipeline on the Honda humanoid robot - ASIMO.

This paper presents an integrative approach to solve the coupled problem of reaching and grasping an object in a cluttered environment with a humanoid robot. While finding an optimal grasp is often treated independently from reaching to the object, in most situations it depends on how the robot can reach a pregrasp pose while avoiding obstacles. We tackle this problem by introducing the concept of task maps which represent the manifold of feasible grasps for an object. Rather than defining a single end-effector goal position, a task map defines a goal hyper volume in the task space. We show how to efficiently learn such maps using the rapidly exploring random tree algorithm. Further, we generalise a previously developed motion optimisation scheme, based on a sequential attractor representation of motion, to cope with such task maps. The optimisation procedure incorporates the robotpsilas redundant whole body controller and uses analytic gradients to jointly optimise the motion costs (including criteria such as collision and joint limit avoidance, energy efficiency, etc.) and the choice of the grasp on the manifold of valid grasps. This leads to a preference of grasps which are easy to reach. The approach is demonstrated in two reach-grasp simulation scenarios with the humanoid robot ASIMO.

We describe a system for visual interaction developed for humanoid robots. It enables the robot to interact with its environment using a smooth whole body motion control driven by stabilized visual targets. Targets are defined as visually extracted "proto-objects" and behavior-relevant object hypotheses and are stabilized by means of a short-term sensory memory. Selection mechanisms are used to switch between behavior alternatives for searching or tracking objects as well as different whole body motion strategies for reaching. The decision between different motion strategies like reaching with right or left hand or with and without walking is made based on internal predictions that use copies of the whole-body control algorithm. The results show robust object tracking and a smooth interaction behavior that includes a large variety of whole-body postures.

We propose a self collision avoidance system that superposes trajectories in order not only to protect the robot's hardware but also to enable continuous motions. The system runs in real-time so that the robot can work in an uncertain environment. It is based on virtual forces between close segments of the robot. The avoidance movements are blended with a whole body motion control in order to change the priority between target reaching and collision avoidance. The blending is performed autonomously without the necessity of external switching. Our method works both while the robot is standing and walking. Reaching motions from the front to the side of the body without the arm colliding with the body are possible. Even if the target is inside the body, the arm stops at the closest point to the target outside of the body. Our method can be used for other applications: We apply it to realizing a "body schema" and for "occlusion avoidance."

This paper presents a whole body motion algorithm and shows some steps towards its feasibility in complex scenarios. We employ the framework of Liegeois, (1977) which solves the redundant inverse kinematics problem on velocity level. To make the controller suitable for a variety of different applications, task descriptors for the relative effector positions as well as a one-and two-dimensional attitude representation are proposed. The inverse kinematics are extended by allowing for "displacement intervals" which are formulated in task space. The proposed control scheme guarantees that the effector motion lies within the specified interval. However, the motion inside the interval is determined by optimization criteria, which can effectively be utilized to generate a more flexible and robust motion. We discuss an example and show simulation and experimental results on the humanoid robot ASIMO

The biped robot "Johnnie" is designed to achieve a dynamically stable gait pattern, allowing for high walking velocities. Besides accurate and fast sensors, a powerful computer system is essential for the performance and stability of the machine. The control system requires hard real-time capabilities and low cycle times. With our new control concept and the new computer system, walking speeds of 2.4 km/h have been achieved in experiments.

The biped robot "Johnnie" is designed to achieve a dynamically stable gait pattern, allowing for high walking velocities. Very accurate and fast sensors were developed for the machine. In particular the design of the 3D-orientation sensor and the 6-axes force-torque sensor are presented. The control scheme is based on the information from these sensors to deal with unstructured terrain and disturbances. Two different implementations are investigated: a computed torque approach and a trajectory control with adaptive trajectories. Walking speeds of 2.0 km/h have been achieved in experiments.

Within the large variety of existing and still newly emerging biped walking machines Jogging JOHNNIE represents a fundamental study on fast walking and on especially adapted foot-dynamics. Air phases are possible with two feet lifted off the ground. Each foot possesses six degrees of freedom with respect to the body and additionally seven degrees of freedom within the local foot environment. The overall system includes 23 degrees of freedom. It is 1.80 m, large and weighs 40 kg. Up to now stable walking has been achieved.

Deals with the design and control of an anthropomorphic autonomous biped robot. The objective is to realize a dynamically stable, three-dimensional walking and jogging motion. The design, sensors and electronics of the robot are introduced. Particular emphasis has been devoted to achieving a high power-to-weight ratio. The corresponding methods for weight reduction are presented. The control scheme is discussed. It is based on the method of feedback linearization employing the equations of motion of the system.

The project of a two-legged walking robot is presented. The control system is based on a nonlinear control scheme using the method of feedback linearization. Special effort has been devoted to the trajectory generation which is computed such that the system remains controllable throughout the entire gait cycle. The controller and simulation algorithm have been implemented parametrically. Parameter studies have been performed which lead to the final design. The mechanical design with particular emphasis on actuators and sensors is presented.