Robotic surgery has reached a high level of maturity and has become an integral part of standard surgical care. However, existing surgeon consoles are bulky, take up valuable space in the operating room, make surgical team coordination challenging, and their proprietary nature makes it difficult to take advantage of recent technological advances, especially in virtual and augmented reality. One potential area for further improvement is the integration of modern sensory gloves into robotic platforms, allowing surgeons to control robotic arms intuitively with their hand movements. We propose one such system that combines an HTC Vive tracker, a Manus Meta Prime 3 XR sensory glove, and SCOPEYE wireless smart glasses. The system controls one arm of a da Vinci surgical robot. In addition to moving the arm, the surgeon can use fingers to control the end-effector of the surgical instrument. Hand gestures are used to implement clutching and similar functions. In particular, we introduce clutching of the instrument orientation, a functionality unavailable in the da Vinci system. The vibrotactile elements of the glove are used to provide feedback to the user when gesture commands are invoked. A qualitative and quantitative evaluation has been conducted that compares the current device with the dVRK console. The system is shown to have excellent tracking accuracy, and the new interface allows surgeons to perform common surgical training tasks with minimal practice efficiently.

Robotic surgery offers enhanced precision and adaptability, paving the way for automation in surgical interventions. Cholecystectomy, the gallbladder removal, is particularly well-suited for automation due to its standardized procedural steps and distinct anatomical boundaries. A key challenge in automating this procedure is dissecting with accuracy and adaptability. This paper presents a vision-based autonomous robotic dissection architecture that integrates real-time segmentation, keypoint detection, grasping and stretching the gallbladder with the left arm, and dissecting with the other arm. We introduce an improved segmentation dataset based on videos of robotic cholecystectomy performed by various surgeons, incorporating a new "liver bed" class to enhance boundary tracking after multiple rounds of dissection. Our system employs state-of-the-art segmentation models and an adaptive boundary extraction method that maintains accuracy despite tissue deformations and visual variations. Moreover, we implemented an automated grasping and pulling strategy to optimize tissue tension before dissection upon our previous work. Ex vivo evaluations on porcine livers demonstrate that our framework significantly improves dissection precision and consistency, marking a step toward fully autonomous robotic cholecystectomy.

The KO (Kronecker Operation) code is a recent deep-learned error-correcting code using a neural network architecture to generalize a Reed-Muller code with Dumer decoding. Analyzing the encoder modules and using ablation techniques, we give interpretations of the KO encoder which significantly reduce the number of parameters. We also discuss interpretability aspects of the KO decoder. The interpretation opens up possibilities to give an explicit representation of KO codes, which could be useful for more efficient learning of KO codes and explaining the learning mechanism underlying the empirical observations made about its performance in previous work.

Several researchers have used deep learning to obtain novel feedback codes. Two such codes for AWGN channels with passive (possibly noisy) output feedback are Deepcode which employs a bit-by-bit rate 1/3 encoder, and Lightcode, which is a symbol-by-symbol code inspired by the Schalkwijk-Kailath (SK) scheme. Here, we build on prior work to interpret these codes by 1) providing the optimal maximum a posteriori (MAP) decoder for our simple non-linear interpretable encoder of a single-bit, two-round code that accurately approximates both single-bit Deepcode and Lightcode. This non-linear interpretable coding scheme, which mimics these codes, turns out to resemble both the functional form and performance of the Polyanskiy-Poor-Verdu (PPV) single bit feedback scheme that minimizes energy transmission asymptotically. 2) We extend our non-linear interpretable code to support more than one bit and two rounds, again providing an optimal MAP decoder. This remarkably simple and power-efficient nonlinear scheme provides insight into Lightcode.

Collaborative manipulation task often requires negotiation using explicit or implicit communication. An important example is determining where to move when the goal destination is not uniquely specified, and who should lead the motion. This work is motivated by the ability of humans to communicate the desired destination of motion through back-and-forth force exchanges. Inherent to these exchanges is also the ability to dynamically assign a role to each participant, either taking the initiative or deferring to the partner’s lead. In this paper, we propose a hierarchical robot control framework that emulates human behavior in communicating a motion destination to a human collaborator and in responding to their actions. At the top level, the controller consists of a set of finite-state machines corresponding to different levels of commitment of the robot to its desired goal configuration. The control architecture is loosely based on the human strategy observed in the human-human experiments, and the key component is a real-time intent recognizer that helps the robot respond to human actions. We describe the details of the control framework, feature engineering and training process of the intent recognition. The proposed controller was implemented on a UR10e robot (Universal Robots) and evaluated through human studies. The experiments show that the robot correctly recognizes and responds to human input, communicates its intent clearly, and resolves conflict. We report success rates and draw comparisons with human-human experiments to demonstrate the effectiveness of the approach.

Deep learning methods have recently been used to construct non-linear codes for the additive white Gaussian noise (AWGN) channel with feedback. However, there is limited understanding of how these black-box-like codes with many learned parameters use feedback. This study aims to uncover the fundamental principles underlying the first deep-learned feedback code, known as Deepcode, which is based on an RNN architecture. Our interpretable model based on Deepcode is built by analyzing the influence length of inputs and approximating the non-linear dynamics of the original black-box RNN encoder. Numerical experiments demonstrate that our interpretable model - which includes both an encoder and a decoder - achieves comparable performance to Deepcode while offering an interpretation of how it employs feedback for error correction.11This work was supported by NSF under awards 1900911, 2217023, and 2240532, and by the AI National Laboratory Program (RRF-2.3.1-21-2022- 00004). The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of the NSF.

As new deep-learned error-correcting codes continue to be introduced, it is important to develop tools to interpret the designed codes and understand the training process. Prior work focusing on the deep-learned TurboAE has both interpreted the learned encoders post-hoc by mapping these onto nearby "interpretable" encoders, and experimentally evaluated the performance of these interpretable encoders with various decoders. Here we look at developing tools for interpreting the training process for deep-learned error-correcting codes, focusing on: 1) using the Goldreich-Levin algorithm to quickly interpret the learned encoder; 2) using Fourier coefficients as a tool for understanding the training dynamics and the loss landscape; 3) reformulating the training loss, the binary cross entropy, by relating it to encoder and decoder parameters, and the bit error rate (BER); 4) using these insights to formulate and study a new training procedure. All tools are demonstrated on TurboAE, but are applicable to other deep-learned forward error correcting codes (without feedback).

This paper proposes a neural network-based user simulator that can provide a multimodal interactive environment for training Reinforcement Learning (RL) agents in collaborative tasks involving multiple modes of communication. The simulator is trained on the existing ELDERLY-AT-HOME corpus and accommodates multiple modalities such as language, pointing gestures, and haptic-ostensive actions. The paper also presents a novel multimodal data augmentation approach, which addresses the challenge of using a limited dataset due to the expensive and time-consuming nature of collecting human demonstrations. Overall, the study highlights the potential for using RL and multimodal user simulators in developing and improving domestic assistive robots.

Deep learning has been used recently to learn error-correcting encoders and decoders which may improve upon previously known codes in certain regimes. The encoders and decoders are learned "black-boxes", and interpreting their behavior is of interest both for further applications and for incorporating this work into coding theory. Understanding these codes provides a compelling case study for Explainable Artificial Intelligence (XAI): since coding theory is a well-developed and quantitative field, the interpretability problems that arise differ from those traditionally considered. We develop post-hoc interpretability techniques to analyze the deep-learned, autoencoder-based encoders of TurboAE-binary codes, using influence heatmaps, mixed integer linear programming (MILP), Fourier analysis, and property testing. We compare the learned, interpretable encoders combined with BCJR decoders to the original black-box code.

Human-human collaboration is characterized by a back-and-forth, where an action of one agent elicits the response of the other. This interaction is inherently multimodal and includes both high-level modalities such as language and low-level ones such as force exchanges. In this work, we investigate human collaborative manipulation: we show distinct patterns that can be identified in low-level physical data and that can be interpreted as primitives used by humans to negotiate about various aspects of the motion and to execute the motion. These primitives provide a high-level interpretation of the interaction and can be used to connect low-level behavior to language. We describe the human study used to collect the data, the data analysis process, and discuss how the identified primitives could be used by a robot’s interaction manager to mediate physical Human-Robot Interaction (pHRI).

In a collaborative task and the interaction that accompanies it, the participants often take on distinct roles, and dynamically switch the roles as the task requires. A domestic assistive robot thus needs to have similar capabilities. Using our previously proposed Multimodal Interaction Manager (MIM) framework, this paper investigates how role switching for a robot can be implemented. It identifies a set of primitive subtasks that encode common interaction patterns observed in our data corpus and that can be used to easily construct complex task models. It also describes an implementation on the NAO robot that, together with our original work, demonstrates that the robot can take on different roles. We provide a detailed analysis of the performance of the system and discuss the challenges that arise when switching roles in human-robot interactions.

With rapid advances in social robotics, humanoids and autonomy, robot assistants appear to be within reach. However, robots are still unable to effectively interact with humans in activities of daily living. One of the challenges is in the frequent use of multiple communication modalities when humans engage in collaborative activities. In this paper, we propose a Multimodal Interaction Manager, a framework for an assistive robot that maintains an active multimodal interaction with a human partner while performing physical collaborative tasks. The heart of our framework is a Hierarchical Bipartite Action-Transition Network (HBATN), which allows the robot to infer the state of the task and the dialogue given spoken utterances and observed pointing gestures from a human partner, and to plan its next actions. Finally, we implemented this framework on a robot to provide preliminary evidence that the robot can successfully participate in a task-oriented multimodal interaction.

In this paper, we focuse on mitigating the damage due to a seismic event in a mid-size building using wireless structure control (WSC). We consider the nature of wireless communication channels and integrates it into a structural control system that minimizes the building response to external excitation - earthquake events. The control and communication constraints in terms of delay and channel reliability are analyzed and incorporated into the regulator model, which incorporates both estimator and feedback command. The short packet formulation provides the ability to link latency and channel reliability, i.e. error rates. This error-delay coupling allows us to analyze the performance of the control system for several scenarios and channel variation models. The simulations show that measurement error harmfully impacts the system performances more than delayed measurements. On the other hand, it is demonstrated that delayed control commands have more impact on the structural response than lost command packets.

Object handover is a common physical interaction between humans. It is thus also of significant interest for human-robot interaction. In this paper, we are focused on robot-to-human object handover. The main challenge in this case is how to reduce the failure rate, i.e., to ensure that the object does not fall (object safety), while at the same time allowing the human to easily acquire the object (smoothness). To endow the robot with a failure recovery mechanism, we investigated how humans detect failure during the transfer phase of the handover. We conducted a human study that showed that a human giver primarily relies on vision rather than haptic sensing to detect the fall of the object. Motivated by this study, a robotic handover system is proposed that consists of a motion sensor attached to the robot's gripper, a force sensor at the base of the gripper, and a controller that is capable of regrasping the object if it starts falling. The proposed system is implemented on a Baxter robot and is shown to achieve a smooth and safe handover.

In robot to human object handover, the configuration (position and orientation) in which the object is transferred should be selected so that the handover is safe and comfortable for the human. The trajectory on which the robot moves the object to the point of transfer should be also selected so that the robot intention is clear and the handover feels natural to the human. In this paper, we propose to select the configuration for the transfer and the trajectory to reach this configuration based on what humans do in human-human handovers. We describe a human study designed to investigate the human-human handover and propose an ergonomic model that can predict object transfer position observed in the study. A human-robot experiment is then conducted that shows that the proposed model generates transfer positions that match the preferred height and distance relative to the human.

Humans rely on a multitude of senses to achieve a safe handover. The challenge for a robot-human handover is to both equip the robot with the appropriate sensors as well as devise handover controllers that effectively use them to prevent failures. In this paper we use object acceleration as an indicator of an impending handover failure and propose a handover controller that has re-grasping mechanism to prevent falling of the object in case of an imperfect handover. The work is motivated by our observation that humans primarily rely on vision to prevent handover failure. We also propose a novel acceleration sensing setup that can be integrated into a robot gripper. The handover controller is implemented on Baxter Research robot equipped with the proposed sensor and is shown to be effective in preventing object fall.

Human grasp has been studied extensively and many taxonomies have been developed to classify different grasp types. In this work, we collect a comprehensive list of grasp types and investigate the force distribution patterns among them. We conduct a human study and collect force data in various grasp types. Data collection is performed by utilizing a data glove that has seventeen force sensors. We analyze the measured forces and identify different patterns in force distribution. Employing voting technique over various clustering methods, we propose a robust clustering for the force patterns, namely the grasp taxonomy in force domain. The proposed grasp taxonomy can be exploited in designing robotic prosthetic hands, designing grasp controllers and action recognition in human-robot interaction.

One of the most challenging aspects of cooperative manipulation is coordination process between haptically-coupled subjects. The interaction force is believed to play a significant role in this process. In this paper, we propose a model for the interaction force and validate our model through a human study. The human study includes both bimanual and dyadic modes in three different scenarios, specifically designed to study the effect of coordination process in performance of the cooperation. We consider five different performance metrics to measure several aspects of the cooperation. The statistical analysis of these performance indexes proves that, while our model can explain the human behavior in different scenarios and different modes, the alternative models fail to do so.

During collaborative object manipulation, the interaction forces provide a communication channel through which humans coordinate their actions. In order for the robots to engage in physical collaboration with humans, it is necessary to understand this coordination process. Unfortunately, there is no intrinsic way to define the interaction forces. In this study, we propose a model that allows us to compute the interaction force during a dyadic cooperative object manipulation task. The model is derived directly from the existing theories on human arm movements. The results of a user study with 22 human subjects prove the validity of the proposed model. The model is then embedded in a control strategy that enables the robot to engage in a cooperative task with a human. The performance evaluation of the controller through simulation shows that the control strategy is a promising candidate for a cooperative human-robot interaction.

While hand trajectory has been successfully modeled for single arm reaching movement, few works have considered the bimanual reaching movement and no study has modeled the dyadic reaching movement. In a bimanual task, both hands belong to the same person, while in a dyadic task each hand belongs to a different person. In this paper, we study both bimanual and dyadic reaching movements and show that the motion trajectory follows the minimum-jerk trajectory. To the best of our knowledge, this is the first work that studies the dyadic reaching movements. Furthermore, we show that our model is consistent with the existing theories on single arm motions, when applied to each of the cooperating arms.

This paper presents an investigation of human physical interaction. In particular, we describe how data from pressure sensors mounted on a glove worn by a human can be mapped to manipulation actions; the actions can in turn be used to interpret physical interaction during elderly care. The work is part of the RoboHelper project, which aims to build a multimodal communication interface for assistive robots for the elderly. Human-human physical interaction during elderly care and in a realistic setting is studied in this work with the aim of using the learned insights to develop corresponding robot interfaces. The contribution of this work is the identification of various types of physical manipulation actions that take place when an elder is assisted in performing activities of daily living in a natural setting. As part of the RoboHelper project, it has been shown that the knowledge of actions involving physical manipulation of objects helps in understanding the spoken language. More specifically, it improves the resolution of third person pronouns/deictic words and the classification of dialogue acts. In this work we show that pressure sensor data can be used to automatically recognize such physical manipulation actions. The automatic recognition of physical manipulation actions may facilitate future studies of multimodal interaction by greatly reducing the time required for manual annotations. It is also useful for learning from demonstration, a popular approach in human-robot interaction research.

Ensuring the correct behavior of cyber physical systems at run time is of critical importance for their safe deployment. Any malfunctioning of such systems should be detected in a timely manner for further actions. This paper addresses the issue of how quickly a monitor raises an alarm after the occurrence of a failure in cyber physical systems. Towards this end, it introduces a class of systems called exponentially converging monitorable systems. The paper shows that failures in these systems can be detected fast by employing the traditional threshold monitors. It shows that the expected failure detection time for exponentially converging monitorable systems has logarithmic relationship with the inverse of the chosen threshold value. The paper identifies well defined natural classes of these systems. Experimental results are presented that confirm the theoretical results on the relationship between the failure detection time and the chosen threshold values.

The lack of haptic feedback during minimally invasive surgery can cause significant tissue damage and increase morbidity. Estimating the applied force from endoscopic images is a promising approach, especially using binocular images. However, many existing operation rooms are only equipped with monocular endoscopes, making force estimation more problematic. In this paper a new method for estimating the applied force from monocular endoscope images is proposed. The main contribution is the concept of virtual template that enables modeling of surface deformation without the knowledge of the undeformed shape. Results of the in vitro experiment with the lamb liver support the practicality and effectiveness of the proposed method.

In this paper we describe our progress towards understanding human communication through physical interaction. We describe a classification algorithm that can recognize four classes of actions that frequently occur during collaborative manipulation of planar objects. These actions were selected based on a user study involving dyads of elderly and care-giver in a realistic setting. Further user studies were conducted to collect the data necessary to develop the classification algorithm. As part of the data collection we also developed a sensory glove. The classification algorithm gives insight into human collaborative manipulation. More precisely, it identifies features in the data that are significant for classification. This information is particularly interesting as it only relies on physical aspects of the interaction and not on any particular sensor. As a result, the described work does not depend on any particular hardware and can be directly used by other researchers in human-robot interaction to develop further experiments and studies.

The transition to Nano-scale devices is expected to open the way for highly complex parallel systems. However, decreased reliability and strict interconnect limitations are two important challenges that the devices need to overcome. To do so, such systems have to adapt to the changing workload and faults, employing decentralized protocols to coordinate among the many components. In this paper, we investigate algorithms for evenly distributing resources among locally connected components, so that the system can dynamically self-balance as the availability of resources changes. We propose two efficient Decentralized Protocols that achieve near-optimal resource distributions. The protocols are scalable and guarantee the desired fair distribution regardless of the interconnect topology.

This paper describes a haptic simulator that has been developed as an aid for the sensorimotor skill acquisition in dentistry. An important feature of the simulator is the ability to generate templates of position and force trajectories for the students to follow. Furthermore, the simulator has a mechanism, haptic playback, to help the students follow and learn these templates. Using this feature, the teacher can perform a procedure in the haptic simulator and record her actions. The trainee is then able to observe the recorded procedure and follow the correct trajectory in the position and force. Several haptic playback techniques are reviewed and the procedure implemented in the simulator is described in detail. We also describe hardware and software components of the simulator and their functionality.We conclude by describing the results of a preliminary classroom evaluation of the simulator.

Optimal control of robotic systems with logical constraints is an instance of a hybrid optimal control problem. It has been traditionally treated as a mixed-integer programming problem (MIP) which is of combinatorial complexity. This paper proposes a new approach for transforming logical constraints into inequality and equality constraints involving only continuous variables. In this way the hybrid optimal control problem is converted to a smooth optimal control problem that can in turn be solved using traditional nonlinear programming methods, thereby dramatically reducing the computational complexity of finding the solution. We illustrate the techniques by solving an optimal path planning problem for multiple unmanned aerial vehicles (UAVs) with collision avoidance. Simulation results are given to show the effectiveness of the approach.

This paper applies the modified Fisher information matrix (FIM) motion algorithm previously proposed by the authors to the task of locating a circular biochemical source. We develop the diffusion model for a circular source and perform control theoretic analysis of the resulting FIM motion algorithm. While in our previous work we established that the source location is an equilibrium point of the system, in the present paper we show that due to the consistency of the maximum likelihood (ML) estimator, the equilibrium point is unique. Simulations are presented that compare our motion algorithm to conventional concentration gradient motion algorithms. The simulations confirm that by using our motion algorithm, the circular biochemical source is located with a high degree of accuracy.

This paper studies the problem of stabilizing wheeled mobile robots (WMRs) subject to wheel slippage to a predefined set. When slippage of the wheels can occur, WMRs can be modeled as hybrid systems. Model predictive control for such systems typically results in numerical methods of combinatorial complexity. We show that recently developed embedding techniques can be used to formulate numerical algorithms for the hybrid model predictive control (MPC) problem that have the same complexity as the MPC for smooth systems. We also discuss in detail the numerical techniques that lead to efficient and robust MPC algorithms. Examples are given to illustrate the effectiveness of the approach.

Existing penalty-based haptic rendering approaches compute penetration depth in strictly translational sense and cannot properly take object rotation into account. We aim to provide a theoretical foundation for computing the penetration depth on the group of rigid-body motions SE(3). We propose a penalty-based six-degrees-of-freedom (6-DOF) haptic rendering algorithm based on determining the closestpoint projection of the inadmissible configuration onto the set of admissible configurations. Energy is used to define the metric on the configuration space. Once the projection is found the 6-DOF wrench can be computed. The configuration space is locally represented with exponential coordinates to make the algorithm more efficient. Numerical examples compare the proposed algorithm with the existing approaches and show its advantages.

The main contribution of this work is a method for robust stabilization of three-dimensional bipedal walking robots with more than one degree of under-actuation. The general framework we previously developed for stabilization of periodic orbits for hybrid systems with impact effects is shown to be applicable to three-dimensional under-actuated bipedal robots. It is shown how periodic solutions for the hybrid dynamical equations describing three-dimensional under-actuated bipedal robots can be found and that these periodic solutions (walking gaits) can be robustly stabilized if a certain semi-definite program can be solved. The fact that the robust control synthesis problem can be cast as a semi-definite program implies that computationally efficient linear matrix inequality (LMI) solvers can be used to find the controllers. We demonstrate the methodology through the simulations on a five-link spatial biped with two degrees of under-actuation

The paper proposes an improved model and its approximation for a diffusion of a biochemical agent in the air. Based on the model, a new motion control algorithm based on the Fisher information matrix (FIM) for detecting and localizing a biochemical source is proposed. We show that the location of the biochemical source is an equilibrium point of the system under such control. Simulations with a single moving sensor show that the biochemical source is located with a high degree of accuracy while the trajectory of the sensor converges to the source

The paper proposes several improvements on the Direction of Gradient (DOG) algorithm proposed in [1] for detecting and localizing a biochemical source with moving sensors. In particular, we show that the DOG algorithm can be turned into a distributed control scheme for a mobile sensing network, and that the maximum likelihood estimation proposed in the original algorithm can be replaced with more computationally efficient numerical procedures. Simulations on a single sensor and on a group of mobile sensors are provided that show that the proposed modifications simplify the original algorithm and improve its performance.

This paper investigates haptic interaction with virtual environments composed of objects with diverse dynamic properties. Passivity is often used for stability analysis of haptic systems. We demonstrate that when the dynamics of the virtual environment during the interaction changes, the approaches using traditional notion of passivity for hybrid systems can be used to design stable interaction strategies for such systems.

The main contribution of this work is a general method for stabilization of periodic orbits for hybrid systems with impact effects. Our primary motivation is controller synthesis for walking robots, but the method can be also applied to problems such as flight control or automotive control. Limit cycles of hybrid systems are characterized by the fact that they span different dynamic regimes. For smooth systems, dynamics of the system along the limit cycle can be decomposed into the transverse and tangential components. We demonstrate that this decomposition can be adapted to hybrid systems. Furthermore, we show that when the transverse dynamics is linearized and discretized, the resulting robust control synthesis problem can be cast as a semidefinite program and thus efficiently solved. We demonstrate our results through the simulation on a simple planar biped robot.

This paper presents a computationally efficient method for deriving coordinate representations for the equations of motion and the affine connection describing a class of Lagrangian systems. We consider mechanical systems endowed with symmetries and subject to nonholonomic constraints and external forces. This method is demonstrated on two robotic locomotion mechanisms known as the snake board and the roller racer. The resulting coordinate representations are compact and lead to straightforward proofs of various controllability results.

In a typical dextrous manipulation task, a goal configuration is reached through a sequence of continuous motions. Most often, a motion plan is computed offline and subsequently used as a reference trajectory for a feedback controller. We present an alternative approach that only relies on feedback, no motion planning is necessary. Different feedback controllers are constructed and composed in such a way that the object moves toward the goal configuration. Switches between the controllers are not planned ahead, they result from the feedback itself. Discrete features such as finger gait are therefore generated online. The approach is based on our previous results on stabilization of hybrid systems.

This paper studies a class of hybrid mechanical systems that locomote by switching between constraints defining different dynamic regimes. We develop a geometric framework for modeling smooth phenomena such as inertial forces, holonomic and nonholonomic constraints, as well as discrete features such as transitions between smooth dynamic regimes through plastic and elastic impacts. We focus on devices that are able to switch between constraints at an arbitrary point in the configuration space. This class of hybrid mechanical control systems can be described in terms of affine connections and jump transition maps that are linear in the velocity. We investigate two notions of local controllability, the equilibrium and kinematic controllability, and provide sufficient conditions for each of them. The tests rely on the assumption of zero velocity switches. We illustrate the modeling framework and the controllability tests on a planar sliding, clamped, and rolling device. In particular, we show how the analysis can be used for motion planning.

We analyze a class of mechanisms that locomote by switching between constraints. Because of the hybrid nature of such systems, most of the existing analysis tools, developed primarily for smooth systems, can not be directly applied. Our aim is to exploit the special structure provided by Lagrangian mechanics to study the controllability of this class of mechanisms. We base the analysis on a series representation of the evolution of the system. Our main result is a description of trajectories involving switches between constraints at nonzero velocity (impacts) in the presence of large inertial forces (drift). The analysis provides a basis for local motion planning. The results are applied to an example of a two-link planar mechanism that can locomote by clamping one of the links.

We present a framework for designing stable control schemes for systems whose dynamics change. The idea is to develop a controller for each of the regions defined by different dynamic characteristics and design a switching scheme that guarantees the stability of the overall system. We derive sufficient conditions for the stability of the switching scheme for systems evolving on a sequence of embedded manifolds. An important feature of the proposed framework is that if the conditions are satisfied by pairs of controllers adjacent in the hierarchy, the overall system will be stable. This makes the application of our results particularly straight forward. The methodology is applied to stabilization of a shimmying wheel, where changes in the dynamic behaviour are due to switches between sliding and rolling.

This paper investigates methods for computing a smooth motion that interpolates a given set of positions and orientations of a rigid body. To make the interpolation independent of the representation of the motion, we use the coordinate-free framework of differential geometry. Inertial and body-fixed reference frames must be chosen to describe the position and orientation of the rigid body. We show that trajectories that are independent of the choice of these frames can be obtained by using the exponential map. Since these trajectories may exhibit rapid changes in velocity or its higher derivatives, a method for finding the maximally smooth interpolating curve is developed. Trajectories computed by both methods are compared on an example.

This paper investigates methods for computing a smooth motion that interpolates a given set of positions and orientations. The position and orientation of a rigid body can be described with an element of the group of spatial rigid body displacements, SE(3). To find a smooth motion that interpolates a given set of positions and orientations is therefore the same as finding an interpolating curve between the corresponding elements of SE(3). To make the interpolation on SE(3) independent of the representation of the group, we use the coordinate-free framework of differential geometry. It is necessary to choose inertial and body-fixed reference frames to describe the position and orientation of the rigid body. We first show that trajectories that are independent of the choice of these frames can be obtained by using the exponential map on SE(3). However, these trajectories may exhibit rapid changes in the velocity or higher derivatives. The second contribution of the paper is a method for finding the maximally smooth interpolating curve. By adapting the techniques of the calculus of variations to SE(3), necessary conditions are derived for motions that are equivalent to cubic splines in the Euclidean space. These necessary conditions result in a boundary value problem with interior-point constraints. A simple and efficient numerical method for finding a solution is then described. Finally, we discuss the dependence of the computed trajectories on the metric on SE(3) and show that independence of the trajectories from the choice of the reference frames can be achieved by using a suitable metric.

This paper addresses the problem of generating smooth trajectories between an initial and a final position and orientation in space. The main idea is to define a functional depending on velocity or its derivatives that measures smoothness of trajectories and find a trajectory that minimizes this functional. In order to ensure that the computed trajectories are independent of the parametrization of positions and orientations, we use the notions of Riemannian metric and covariant derivative from differential geometry and formulate the problem as a variational problem on the Lie group of spatial rigid body displacements. We show that by choosing an appropriate measure of smoothness, the trajectories can be made to satisfy boundary conditions on the velocities or higher order derivatives. Dynamically smooth trajectories can be obtained by incorporating the inertia of the system into the definition of the Riemannian metric. We state the necessary conditions for the shortest distance, minimum acceleration and minimum jerk trajectories.

We study the 6/spl times/6 Cartesian stiffness matrix. We show that the stiffness of a rigid body subjected to conservative forces and moments is described by a (0,2) tensor which is the Hessian of the potential function. The key observation of the paper is that since the Hessian depends on the choice of an affine connection in the task space, so will the Cartesian stiffness matrix. Further, the symmetry of the Hessian and thus of the stiffness matrix depends on the symmetry of the connection. The connection that is implicit in the definition of the Cartesian stiffness matrix through the joint stiffness matrix (Salisbury, 1980) is made explicit and shown to be symmetric. In contrast, the direct definition of the Cartesian stiffness matrix in Griffis (1993), Ciblak and Lipkin (1994) and Howard et al. (1996) is shown to be derived from an asymmetric connection. A numerical example is provided to illustrate the main ideas of the paper.

This paper studies human dual arm manipulation tasks and develops a computational model that predicts the trajectories and the force distribution for the coordination of two arms moving an object between two given positions and orientations in a horizontal plane. Our ultimate goal is to understand the dynamics of dual arm coordination in order to develop better robot control algorithms. Our computational model is based on the hypothesis proposed by Uno et al. (1989) who suggest that human movements minimize the integral of the norm of the rate of change of actuator torques. We compare the experimental trajectories and force distributions with this computational model. The first important observation is that the trajectories show a significant degree of repeatability across trials and across subjects. Next, we observe that the trajectories in the sagittal and frontal plane are characterized by asymmetric features that are hard to model using such integral cost functions. Finally, we show that the internal forces play an important role in trajectory generation. While these are repeatable across trials, they vary significantly from subject to subject.

We address the motion planning for "fixtureless" material-handling with multiple manipulators on nonholonomic carts. The mobile manipulators possess the ability to manipulate and transport objects while holding them in a stable grasp. We present a general approach that allows generation of optimal trajectories and actuator inputs for any given maneuver. Constraints such as limitations on the turning radii of the mobile manipulators or bounds on their separation can be easily incorporated into the planning scheme. Numerical solutions for several maneuvers including abrupt turns, parallel parking in cluttered environments and changes in formation are computed. Finally, we present experimental results with two mobile manipulators.

This paper addresses the general problem of generating smooth trajectories between an initial and a final position and orientation. A functional depending an velocity and its higher derivatives involving a left invariant Riemannian metric on SE(3) is used to measure the smoothness of a trajectory. The problem of determining a smooth trajectory between two points is formulated as a variational problem on SE(3). The authors derive necessary conditions for the shortest distance and minimum jerk trajectories and solve the resulting two-point boundary value problem.

We present a kinematic model of a paraplegic subject walking with crutches where the subject with the crutches is modeled as a parallel kinematic structure. The model is employed to investigate if certain quadrupedal gait patterns can be implemented with functional electrical stimulation. The study is motivated by the fact that the existing crutch-assisted gait realized by the electrical stimulation is slow and energy inefficient. Gait patterns that would improve the walking are identified. The main characteristic of the patterns is that some of their states are not statically stable. During such states, the subject is supported by only a leg and a crutch. It is demonstrated that if the forward motion is provided by the stimulation of the plantar flexors the trajectory of the center of the body can closely follow the trajectory that is observed during walking of healthy subjects. We argue that the resulting gait is smooth and energy efficient. In addition, the unstable states make the walking faster.

Problems in robotics and biomechanics such as trajectory planning or resolution of redundancy can be effectively solved using optimal control. Such systems are often subject to unilateral constraints. Examples include tasks involving contacts (e.g., walking, running, multifingered or multiarm manipulation), and other tasks that may not involve contacts but in which the system state or the inputs must satisfy inequality conditions (e.g., limits on actuator forces). This paper shows how problems of optimal control in robotics that involve unilateral constraints can be efficiently solved by first formulating the constrained optimal control problem as an unconstrained problem of the calculus of variations and then solving it using an integral formulation. This method has several advantages over the Pontryagin minimum principle which is traditionally employed to solve such problems. An example of two-arm manipulation with inequality constraints due to Coulomb friction is used to demonstrate the formulation of the problem and the algorithms.

Improved walking of completely paralyzed paraplegic subjects assisted by multichannel functional electrical stimulation (FES) and crutches is proposed. In the resulting quadrupedal gait the center of body (COB) is both actively and passively transferred in the direction of progression. Active transfer of the COB is accomplished by electrical stimulation of ankle plantarflexors of the trailing leg. Passive displacement of the COB occurs during the unstable state when the paraplegic person is only supported by a single foot and the contralateral crutch. Comparison of the time parameters measured during human crawling on knees and arms with the crutch assisted walking is presented. Results of kinematic modeling of the present and the improved FES and crutch assisted walking are displayed in the form of supporting polygons with the trajectory of the COB projected to the ground.

This paper addresses determination of trajectories and force distribution for cooperative manipulation with two arms through optimizing an integral cost function that depends an the actuator forces. We compare the calculated trajectories with the measurements on human subjects performing planar manipulation tasks. Our findings suggest that the trajectories and forces used by humans can be predicted by minimizing the integral of the rate of change of actuator torques over the trajectory. Good match is shown for a class of manipulation tasks in which the person-to-person variability is small. The theoretical foundation for computing the optimal solutions is briefly presented and the advantages of using such schemes for robotic systems are discussed.

The optimization of trajectories and actuator torques for a dual arm manipulation system is considered. Given the starting and final configurations, we find the trajectories that minimize: (a) the integral of the norm of the vector of derivatives of the actuator forces; and (b) the integral of the norm of the actuator forces. In this way both kinematic and actuator redundancy are resolved. The optimization problem reduces to solving a two-point boundary valve problem for coupled, nonlinear differential equations. The effect of different parameters such as preload and inertia are investigated and the results are compared with those obtained using other well-known cost functions.>