Control and Learning of Compliant Manipulation Skills

Humans demonstrate an impressive capability to manipulate fragile objects without damaging them, graciously controlling the force and position of hands or tools. Traditionally, robotics has favored position control over force control to produce fast, accurate and repeatable motion. For extending the applicability of robotic manipulators outside the strictly controlled environments of industrial work cells, position control is inadequate. Tasks that involve contact with objects whose positions are not known with perfect certainty require a controller that regulates the relationship between positional deviations and forces on the robot. This problem is formalized in the impedance control framework, which focuses the robot control problem on the interaction between the robot and its environment. By adjusting the impedance parameters, the behavior of the robot can be adapted to the need of the task. However, it is often difficult to specify formally how the impedance should vary for best performance. Furthermore, fast it can be shown that careless variation of the impedance can lead to unstable regulation or tracking even in free motion. In the first part of the thesis, the problem of how to define a varying impedance for a task is addressed. A haptic human-robot interface that allows a human supervisor to teach impedance variations by physically interacting with the robot during task execution is introduced. It is shown that the interface can be used to enhance the performance in several manipulation tasks. Then, the problem of stable control with varying impedance is addressed. Along with a theoretical discussion on this topic, a sufficient condition for stable varying stiffness and damping is provided. In the second part of the thesis, we explore more complex manipulation scenarios via online generation of the robot trajectory. This is done along two axes 1) learning how to react to contact forces in insertion tasks which are crucial for assembly operations and 2) autonomous Dynamical Systems (DS) for motion representation with the capability to encode a family of trajectories rather than a fixed, time-dependent reference. A novel framework for task representation using DS is introduced, termed Locally Modulated Dynamical Systems (LMDS). LMDS differs from existing DS estimation algorithms in that it supports non-parametric and incremental learning all the while guaranteeing that the resulting DS is globally stable at an attractor point. To combine the advantages of DS motion generation with impedance control, a novel controller for tasks described by first order DS is proposed. The controller is passive, and has the properties of an impedance controller with the added flexibility of a DS motion representation instead of a time-indexed trajectory.

From High-Level to Low-Level Robot Learning of Complex Tasks: Leveraging Priors, Metrics and Dynamical Systems

Nadia Barbara Figueroa Fernandez

Humans have a remarkable way of learning, adapting and mastering new manipulation tasks. With the current advances in Machine Learning (ML), the promise of having robots with such capabilities seems to be on the cusp of reality. Transferring human-level skills to robots, however, is complicated as they involve a level of complexity that cannot be tackled by classical ML methods in an unsupervised way. Such complexities involve: (i) automatically decomposing tasks into control-oriented encodings, (ii) extracting invariances and handling idiosyncrasies of data acquired from human demonstrations, and (iii) learning models that guarantee stability and convergence. In this thesis, we push the boundaries in the learning from demonstration (LfD) domain by addressing these challenges with minimal human intervention and parameter tuning. We introduce novel learning approaches based on Bayesian non-parametrics and kernel-methods while encoding novel metrics and priors, leveraging them with dynamical systems (DS) theory. In the first part of this thesis we focus on learning complex sequential manipulation tasks from heterogeneous and unstructured demonstrations. Such tasks are those composed of a sequence of discrete actions. The particular challenge is learning these tasks without any prior knowledge on the number of actions (unstructuredness) or restrictions as to how the human is demonstrating the task (heterogeneous), e.g. changes in reference frames. We propose a Bayesian non-parametric learning framework that can jointly segment and discover unique discrete actions (and their sequence) in continuous demonstrations of the task. Hence, we learn an entire task from a continuous, unrestricted natural demonstration. The learned structure of the complex tasks and the segmented data were then used to parametrize a hybrid controller to execute two cooking tasks of increasing complexity: (i) single-arm pizza dough rolling, and (ii) a dual-arm vegetable peeling task. Throughout this thesis, we assume that both the human and robot motions are driven by autonomous state-dependent DS. Hence, in the second part of this thesis we offer two novel DS formulations to represent and execute a complex task. We begin by proposing a DS-based motion generator formulation and learning scheme that is capable of automatically encoding continuous and complex motions while ensuring global asymptotic stability. The type of tasks that can be learned with this approach are unparalleled to previous work in DS-based LfD and are validated on production line and household activities. Further, we propose a novel DS formulation and learning scheme that can encode both complex motions and varying impedance requirements along the task; i.e., the robot must be compliant in some regions of the task, while stiff in others. This approach is validated on trajectory tracking tasks where a robot arm must precisely draw letters on a surface. Due to the generalization power and straight-forward learning schemes of the proposed DS-based motion generators, we also apply them to more complex application. Such applications include providing adaptive (i) navigation strategies for mobile agents, and (ii) locomotion and co-manipulation tasks of biped robots. These applications are particularly novel in the LfD domain, as most work is solely focused on robotic arm control. In the last part of the thesis, we explore learning complex behaviors in joint space for single and multi-arm systems.

EPFL2019

Wearable Technologies for Embodied Human-Robot Interaction

Carine Rognon

Robotic teleoperation is fundamental to augment the resilience, precision, and force of robots with the cognition of the operator. However, current interfaces, such as joysticks and remote controllers, are often complicated to handle since they require cognitive effort and learned skills. Wearable interfaces can enable more natural and intuitive interactions with robots, which would make robotic teleoperation accessible to a larger population of users for demanding tasks, such as manipulation or search-and-rescue. The aim of this thesis is to explore solutions to simplify our interactions with robots promoting their access to a broader range of the population. To achieve this, we are presenting a soft upper body exoskeleton, called the FlyJacket, for the bidirectional control of drones. Drones can greatly benefit us as they extend our perception and range of action. The exoskeleton controls a drone by recording torso movement and, through embedded haptic feedback devices, renders either kinesthetic guidance to improve the flight performance or tactile feedback to render the sensation of flying. We developed and tested an interface to control both a simulated and a real drone. The FlyJacket is a soft exoskeleton with arm support conceived to address the challenges of adapting to different morphologies and supporting the user during flight to prevent fatigue. We demonstrated that this novel interface allowed more consistent performance than when performing the same task with a remote controller and users felt more immersed into the flight. Interacting with a robot can be greatly enhanced by having multiple channels of sensorial feedback to increase the awareness of the operator. Information on the state of the drone can be intuitively rendered with haptic feedback. To create this bidirectional interaction with the drone, the two types of haptic feedback - kinesthetic and tactile - have been explored. Kinesthetic feedback was implemented with a cable-driven system to give guidance to the userâs torso position. Performing user studies, we could determine that the embedded guidance improved the flight performance and that a quadratically shaped force feedback curve was the most adequate profile to guide the user. We also established the minimal force difference, defined the perceived magnitude of this system and studied the learning process of users. Tactile feedback was investigated to render the sensation of flying by enhancing flight awareness, realism and immersion. To this end, we developed and embedded a new type of soft actuator that was compliant and lightweight such that it remained wearable and portable. Four devices, placed on the torso, provided feedback by compressing closed air pouches against the skin rendering the sensation of air pressure. A mechanical model and simulation of the pouch device were developed to determine appropriate parameters. We evaluated whether it conveyed useful information to the user and whether it enhanced the experience of flying. We demonstrated that users were able to understand the direction of the cues without prompting, could distinguish the cues quickly, and do so with high accuracy. The device was also used in a simulated flight task and users indicated that it increased the flight realism. We believe that the contributions of this thesis provides insights to the design of intuitive interfaces for human-robot interaction and increases their accessibility to a wider range of the population.

EPFL2019

Learning From Demonstration and Interactive Control of Variable-Impedance to Cut Soft Tissues

Aude Billard, Rui Wu

In this article, we propose an approach to extract variable-impedance during cutting tasks from human demonstrations, so as to ease soft-tissue cutting by robots. We model the dynamic adjustment of the human arm during interactions with the tissue and transfer these adaptive capabilities to the robot, by learning both the motion and change of impedance. To improve performance during task execution, our variable-impedance skill-transfer framework combines the learned model with an interactive-operation and feedback controller. To offer the flexibility of modifying the trajectory at run time, we use a control law based on dynamical systems. We couple this control law with virtual dynamics that describes the cutting dynamics. This ensures that the robot can control the interaction force and can plan the trajectory. The approach is validated in a real robot cutting-experiment of cutting different hardness tissues. Results show that our learning-based feedback controller, with the assistance from interactive-operation models, can effectively improve the task’s success rate, cutting length, cutting depth, and other indicators, as compared to using a fixed and variable-impedance gain controller.

2021

From High-Level to Low-Level Robot Learning of Complex Tasks: Leveraging Priors, Metrics and Dynamical Systems

Nadia Barbara Figueroa Fernandez

EPFL2019

Wearable Technologies for Embodied Human-Robot Interaction

Carine Rognon

EPFL2019