A5069603317- Robotic Locomotion and Control
- Prosthetics and Rehabilitation Robotics
- Muscle activation and electromyography studies
- Modular Robots and Swarm Intelligence
- Biomimetic flight and propulsion mechanisms
- Robot Manipulation and Learning
- Zebrafish Biomedical Research Applications
- Neural dynamics and brain function
- Soft Robotics and Applications
- Motor Control and Adaptation
- Stroke Rehabilitation and Recovery
- Reinforcement Learning in Robotics
- Robotic Path Planning Algorithms
- Micro and Nano Robotics
- Advanced Memory and Neural Computing
- Bat Biology and Ecology Studies
- Balance, Gait, and Falls Prevention
- Underwater Vehicles and Communication Systems
- Robotic Mechanisms and Dynamics
- Spinal Cord Injury Research
- Distributed Control Multi-Agent Systems
- Neural Networks and Applications
- Advanced Materials and Mechanics
- Neurobiology and Insect Physiology Research
- Evolutionary Algorithms and Applications
École Polytechnique Fédérale de Lausanne
2016-2025
University of Tehran
2017
Stanford University
2016
Tohoku University
2016
Arizona State University
2016
Swiss National Science Foundation
2015
Instituto Superior Técnico
2013
École Polytechnique
2007-2011
École Normale Supérieure - PSL
2008-2010
Inspire
2004-2009
Nonlinear dynamical systems have been used in many disciplines to model complex behaviors, including biological motor control, robotics, perception, economics, traffic prediction, and neuroscience. While often the unexpected emergent behavior of nonlinear is focus investigations, it equal importance create goal-directed (e.g., stable locomotion from a system coupled oscillators under perceptual guidance). Modeling with is, however, rather difficult due parameter sensitivity these systems,...
The transition from aquatic to terrestrial locomotion was a key development in vertebrate evolution. We present spinal cord model and its implementation an amphibious salamander robot that demonstrates how primitive neural circuit for swimming can be extended by phylogenetically more recent limb oscillatory centers explain the ability of salamanders switch between walking. suggests mechanisms modulation velocity, direction, type gait are relevant all tetrapods. It predicts have lower...
Presents an approach to movement planning, on-line trajectory modification, and imitation learning by representing plans based on a set of nonlinear differential equations with well-defined attractor dynamics. The resultant plan remains autonomous that forms control policy (CP) which is robust strong external perturbations can be modified additional perceptual variables. We evaluate the system humanoid robot simulation actual robot. Experiments are presented for three types movements:...
Movement imitation requires a complex set of mechanisms that map an observed movement teacher onto one's own apparatus. Relevant problems include recognition, pose estimation, tracking, body correspondence, coordinate transformation from external to egocentric space, matching against previously learned movement, resolution redundant degrees-of-freedom are unconstrained by the observation, suitable representations for imitation, modularization motor control, etc. All these topics themselves...
We present the design of a novel compliant quadruped robot, called Cheetah-cub, and series locomotion experiments with fast trotting gaits. The robot’s leg configuration is based on spring-loaded, pantograph mechanism multiple segments. A dedicated open-loop controller was derived implemented. Experiments were run in simulation hardware flat terrain step down, demonstrating self-stabilizing properties. robot reached running trot short flight phases maximum Froude number FR = 1.30, or 6.9...
An important problem in the control of locomotion robots with multiple degrees freedom (e.g., biomimetic robots) is to adapt locomotor patterns properties environment. This article addresses this for an amphibious snake robot, and aims at identifying fast swimming crawling gaits a variety environments. Our approach uses controller based on biological concept central pattern generators (CPGs) together gradient-free optimization method, Powell's method. A key aspect our that are optimized...
The control of compliant robots is, due to their often nonlinear and complex dynamics, inherently difficult. vision morphological computation proposes view these aspects not only as problems, but rather also parts the solution. Non-rigid body are seen anymore imperfect realizations rigid parts, potential computational resources. applicability this has already been demonstrated for a variety robot problems. Nevertheless, theoretical basis understanding capabilities limitations missing so far....
Abstract Robotic technologies, whether they are remotely operated vehicles, autonomous agents, assistive devices, or novel control interfaces, offer many promising capabilities for deployment in real‐world environments. Postdisaster scenarios a particularly relevant target applying such due to the challenging conditions faced by rescue workers and possibility increase their efficacy while decreasing risks face. However, field‐deployable technologies work have requirements robustness, speed,...
In this paper, we present Salamandra robotica II: an amphibious salamander robot that is able to walk and swim. The has four legs actuated spine allow it perform anguilliform swimming in water walking on the ground. paper first presents new hardware design, which improved version of I. We then address several questions related body-limb coordination robots animals have a sprawling posture like salamanders lizards, as opposed erect mammals (e.g., cats dogs). particular, investigate how speed...
The framework of dynamic movement primitives (DMPs) contains many favorable properties for the execution robotic trajectories, such as indirect dependence on time, response to perturbations, and ability easily modulate given but in its original form remains constrained kinematic aspect movement. In this paper, we bridge gap behavior by extending with force/torque feedback. We propose evaluate a modulation approach that allows interaction objects environment. Through proposed coupling...
The development of robotic cognition and the advancement understanding human form two current greatest challenges in robotics neuroscience, respectively. RobotCub project aims to develop an embodied child (iCub) with physical (height 90 cm mass less than 23 kg) ultimately cognitive abilities a 2.5-year-old child. iCub will be freely available open system which can used by scientists all cognate disciplines from developmental psychology epigenetic enhance systems through study development....
We present a system of coupled nonlinear oscillators to be used as programmable central pattern generators, and apply it control the locomotion humanoid robot. Central generators are biological neural networks that can produce coordinated multidimensional rhythmic signals, under simple input signals. They found both in vertebrate invertebrate animals for locomotion. In this article, we novel composed adaptive learn arbitrary signals supervised learning framework. Using rules implemented...