Humanoid robots are designed for accomplishing a wide variety of tasks in human friendly environments but have redundant many degree-of-freedoms, which makes real-time planning and control extremely challenging. We explore efficient end-pose and motion planning methods for humanoid robots to accomplish tasks in different scenarios, especially in complex and changing environments.
Yiming Yang, Vladimir Ivan, Zhibin Li, Maurice Fallon, Sethu Vijayakumar, iDRM: Humanoid Motion Planning with Real-Time End-Pose Selection in Complex Environments. IEEE International Conf. on Humanoid Robots (Humanoids 2016), Cancun, Mexico (2016). [video]
Yiming Yang, Vladimir Ivan, Wolfgang Merkt, Sethu Vijayakumar, Scaling Sampling–based Motion Planning to Humanoid Robots. IEEE International Conf. on Robotics and Biomimetics (ROBIO 2016), Qingdao, China (2016). [video]
Motion Adaptation in Dynamic Environments
Reacting to environment changes is a big challenge for real world robot applications. We explore novel approaches that allow the robot to quickly adapt to changes, particularly in the presence of moving targets and dynamic obstacles. Typically, a configuration space re-planning or adaptation is required if the environment is changed.
Yiming Yang, Vladimir Ivan and Sethu Vijayakumar, Real-Time Motion Adaptation using Relative Distance Space Representation, Proc. 17th IEEE International Conf. on Advanced Robotics (ICAR 2015), Istanbul, Turkey (2015). [video]
Real-time multi-modal prosthetic hand control
We develop methods to improve the capability of machine learning-based upper-limb prostheses. Our work is focused on exploring multi-modal sensing for performance improvement, as well as the use of regression methods for developing the next-generation, proportionally controlled prosthetic hands.
Kyranou I, Krasoulis A, Erden M S, Nazarapour K, Vijayakumar S, `Real-time classification of multi-modal sensory data for prosthetic hand control`, in Biomedical Robotics and Biomechatronics (BioRob), 2016 6th IEEE International Conference, IEEE, pp. 536-541. [pdf]
Krasoulis, A, Vijayakumar, S & Nazarpour, K 2015, 'Evaluation of regression methods for the continuous decoding of finger movement from surface EMG and accelerometry'. in Neural Engineering (NER), 2015 7th International IEEE/EMBS Conference on. IEEE, pp. 631-634., [pdf]
Krasoulis, A, Nazarpour, K & Vijayakumar, S 2015, 'Towards Low-Dimensionsal Proportional Myoelectric Control'. in Engineering in Medicine and Biology Society (EMBC), 2015 37th Annual International Conference of the IEEE. IEEE, pp. 7155 - 7158. [pdf]
Topology-based Representations for Motion Planning
We have developed methods to combine and exploit different representations for motion synthesis, with specific emphasis on generalization of motion to novel situations. We have demonstrated the benefits of our methods on problems where direct path finding in joint configuration space is extremely hard whereas local optimal control exploiting a representation with different topology can efficiently find optimal trajectories. Further, we have illustrated the successful online motion generalization to dynamic environments on challenging, real world problems.
Vladimir Ivan, Dmitry Zarubin, Marc Toussaint, Taku Komura, Sethu Vijayakumar. Topology-based Representations for Motion Planning and Generalisation in Dynamic Environments with Interactions, IJRR, 2013. [pdf]
Dmitry Zarubin, Vladimir Ivan, Marc Toussaint, Taku Komura and Sethu Vijayakumar. Heirachical Motion Planning in Topological Representations. Proc. Robotics: Science and Systems (R:SS 2012), Sydney, Australia (2012). [pdf]
Vladimir Ivan and Sethu Vijayakumar, Space Time Area Coverage for Robot Motion Synthesis, International Conference on Advanced Robotics (ICAR), 2015. [pdf] [video]
Peter Sandilands, Vladimir Ivan, Taku Komura and Sethu Vijayakumar, Dexterous Reaching, Grasp Transfer and Planning Using Electrostatic Representations, Proc. 2013 IEEE-RAS International Conference on Humanoid Robots, Atlanta, USA (2013). [pdf]
BLUE: A Bipedal Robot with Variable Stiffness and Damping
We have designed a planar bipedal robot with joints capable of physically varying both their stiffness and damping independently – the first of its kind. Informed by human biophysics and locomotion studies, we designed an appropriate (heterogenous) impedance modulation mechanism that fits the necessary torque and stiffness range and rate requirements at each joint while ensuring the right form factor. In addition to hip, knee and ankle, the constructed robot is also equipped with a three part compliant foot modelled on human morphology.
Alexander Enoch, Andrius Sutas, Shinichiro Nakaoka and Sethu Vijayakumar. BLUE: A Bipedal Robot with Variable Stiffness and Damping. Proc. 12th IEEE-RAS International Conference on Humanoid Robots, Osaka, Japan (2012). [pdf]
Alexander Enoch and Sethu Vijayakumar. Rapid Manufacture of Novel Variable Impedance Robots. ASME Journal of Mechanisms and Robotics (2015). [doi]
Optimal Variable Impedance Control in Dynamic Movement Tasks
We have developed techniques that can efficiently compute optimised spatiotemporal modulation of torque and impedance profiles for highly dynamic movements in compliantly actuated robots. The proposed methodology is applied to a ball throwing task where we demonstrate that: (i) the method is able to tailor impedance strategies to specific task objectives and system dynamics, (ii) the ability to vary stiffness leads to better performance in this class of movements, (iii) in systems with variable physical compliance, our methodology is able to exploit the energy storage capabilities of the actuators.
David Braun, Florian Petit, Felix Huber, Sami Haddadin, Patrick van der Smagt, Alin Albu-Schäffer and Sethu Vijayakumar, Robots Driven by Compliant Actuators: Optimal Control under Actuation Constraints,IEEE Transactions on Robotics (IEEE T-RO), 29(5), pp. 1085-1101 (2013). [pdf]
David Braun, Matthew Howard and Sethu Vijayakumar, Optimal Variable Stiffness Control: Formulation and Application to Explosive Movement Tasks, Autonomous Robots, vol. 33, pp. 237-253 (2012) [pdf][DOI]
David Braun, Florian Petit, Felix Huber, Sami Haddadin, Patrick van der Smagt, Alin Albu-Schaeffer and Sethu Vijayakumar. Optimal Torque and Stiffness Control in Compliantly Actuated Robots. Proc. IEEE/RSJ Intl. Conf. on Intelligent Robots and Systems (IROS 2012), Portugal (2012). [pdf]
David Braun, Matthew Howard and Sethu Vijayakumar. Exploiting Variable Stiffness in Explosive Movement Tasks. Proc. Robotics: Science and Systems (R:SS 2011), Los Angeles, CA, USA (2011). [pdf] http://youtu.be/LSU_bdHdMXs
Matthew Howard, David Braun and Sethu Vijayakumar. Constraint-based Equilibrium and Stiffness Control of Variable Stiffness Actuators. Proc. IEEE International Conference on Robotics and Automation (ICRA 2011), Shanghai, China (2011). [pdf]
Andreea Radulescu, Matthew Howard, David Braun and Sethu Vijayakumar. Exploiting Variable Physical Damping in Rapid Movement Tasks. Proc. 2012 IEEE ASME International Conference on Advanced Intelligent Mechatronics, Taiwan (2012). [pdf] AIM 2012 Best Student Paper Award Finalist http://youtu.be/w4xg6mwoLlI
Spatio-temporal Optimization of Multi-phase Movements: Dealing with Contacts and Switching Dynamics
We address the optimal control problem of robotic systems with variable stiffness actuation (VSA) including switching dynamics and discontinuous state transitions. Our focus in this paper is to consider tasks that have multiple phases of movement, contacts and impacts with the environment. By modelling such tasks as an approximate hybrid dynamical system with time-based switching, we develop a systematic methodology to simultaneously optimize control commands, stiffness profiles and temporal aspect of the movement such as switching instances and total movement duration. Numerical evaluations on a simple switching system, a realistic brachiating robot model with VSA, and a hopper with variable stiffness springs demonstrate the effectiveness of the proposed approach.
Konrad Rawlik, Marc Toussaint and Sethu Vijayakumar, An Approximate Inference Approach to Temporal Optimization in Optimal Control, Proc. Advances in Neural Information Processing Systems (NIPS '10), Vancouver, Canada (2010).[pdf]
Jun Nakanishi and Sethu Vijayakumar. Exploiting Passive Dynamics with Variable Stiffness Actuation in Robot Brachiation. Proc. Robotics: Science and Systems (R:SS 2012), Sydney, Australia (2012). [pdf][video]
Jun Nakanishi, Konrad Rawlik and Sethu Vijayakumar. Stiffness and Temporal Optimization in Periodic Movements: An Optimal Control Approach. Proc. IEEE/RSJ Intl. Conf. on Intelligent Robots and Systems (IROS 2011), San Francisco (2011). [pdf]
Andreea Radulescu, Jun Nakanishi and Sethu Vijayakumar, Optimal Control of Multiphase Movements with Learned Dynamics, A. Gruca et al. (eds.), Man–Machine Interactions 4, Advances in Intelligent Systems and Computing 391, pp. 61-76 (2016).[pdf]
Tactile Sensing: In Robots and Humans
Making sense of high dimensional tactile sensor data in robots and humans is an exciting challenge. We have developed a framework that uses active learning to help with sequential gathering of most informative data samples. We use information theoretic criteria to find the optimal actions to estimate parameters that affect the dynamics of objects—such as viscosity or internal degrees of freedom. We have also worked on information encoding in human tactile processing.
Hannes Saal, Jo-Anne-Ting and Sethu Vijayakumar. Active Estimation of Object Dynamics Parameters with Tactile Sensors. Proc. IEEE/RSJ Intl. Conf. on Intelligent Robots and Systems (IROS 2010), Taiwan (2010). [pdf]
Hannes Saal, Jo-Anne-Ting and Sethu Vijayakumar. Active sequential learning with tactile feedback. In: Teh YW and Titterington M (Eds.), Proc. 13th Int. Conf. on Artificial Intelligence and Statistics (AISTATS 2010), JMLR: W&CP 9:677-684, Chia Laguna, Sardinia, Italy (2010). [pdf]
Hannes Saal, Jo-Anne-Ting and Sethu Vijayakumar. Active Filtering for Robot Tactile Learning. In: Workshop on Adaptive Sensing, Active Learning, and Experimental Design, Neural Information Processing Systems (NIPS 2009), Whistler, Canada (2009). [pdf]
Hannes Saal, Sethu Vijayakumar and Roland Johansson. Information about Complex Fingertip Parameters in Individual Human Tactile Afferent neurons. The Journal of Neuroscience, 29(25):8022-8031, (2009). [pdf]
Hannes Saal, Sethu Vijayakumar and Roland Johansson. Information about present and past stimulus features in human tactile afferents. Proc. Computational and Systems Neuroscience COSYNE '08, Salt Lake City, Utah (2008). [poster]
iLIMB Vibrotactile Feedback
In collaboration with prosthesis developer Touch Bionics, we fit subjects with the i-limb, a state-of-the-art prosthetic hand, and an array of vibrating motors to communicate feedback from force and position sensors to the wearer. We have developed a novel manipulandum for understanding the sensorimotor processes involved in object grasping together with a closed-loop prosthetic hand, with 2 degrees of control and 32 channels of vibrotactile feedback of fingertip forces and finger positions.
Ian Saunders and Sethu Vijayakumar. Continuous Evolution of Statistical Estimators for Optimal Decision-Making. PLoS ONE, vol. 7, No. 6 (2012) [pdf]
Ian Saunders and Sethu Vijayakumar. The Role of Feed-Forward and Feedback Processes for Closed-Loop Prosthesis Control. Journal of Neuroengineering and Rehabilitation (JNER), 8:60 (2011). [pdf]
Ian Saunders and Sethu Vijayakumar. A Closed Loop Prosthetic Hand as a Model Sensorimotor Circuit. Proc. ESF Intl. Workshop on Computational Principles of Sensorimotor Learning, Irsee, Germany (2009). [pdf]
Dimensionality Reduction to Exploit Constraints in Reinforcement Learning
We have developed methods to incorporate prior knowledge from demonstrations of individual robot postures into learning by extracting the inherent problem structure to find an efficient state representation. In particular, we used probabilistic, nonlinear dimensionality reduction to capture latent constraints present in the data. By learning policies in the extracted latent space, we were able to solve the planning problem in a reduced space that automatically satisfies task constraints.
Sebastian Bitzer, Matthew Howard and Sethu Vijayakumar. Using Dimensionality Reduction to Exploit Constraints in Reinforcement Learning. Proc. IEEE/RSJ Intl. Conf. on Intelligent Robots and Systems (IROS 2010), Taiwan (2010). [pdf]
Sebastian Bitzer, Stefan Klanke and Sethu Vijayakumar. Does Dimensionality Reduction improve the Quality of Motion Interpolation? Proc. 17th European Symposium on Artificial Neural Networks (ESANN ’09), Bruges, Belgium (2009). [pdf]
Sebastian Bitzer, Ioannis Havoutis and Sethu Vijayakumar. Synthesising Novel Movements through Latent Space Modulation of Scalable Control Policies. Asada et. al (eds.) Proc. Tenth International Conference on the Simulation of Adaptive Behavior (SAB '08), Springer-Verlag LNAI 5040, pp. 199-209, Osaka, Japan (2008). [pdf]
Learning Policies from Variable Constraint Data
We have developed a novel approach for learning (unconstrained) control policies from movement data, where observations come from movements under different constraints. We demonstrate our approach on systems of varying complexity, including kinematic data from the ASIMO humanoid robot with 27 degrees of freedom, and present results for learning from human demonstration.
Matthew Howard, Stefan Klanke, Michael Gienger, Christian Goerick and Sethu Vijayakumar. Methods for Learning Control Policies from Variable Constraint Demonstrations. In: O. Sigaud and J. Peters (eds.): From Motor Learning to Interaction Learning in Robots, SCI 264, pp. 253-291, Springer-Verlag (2010). [pdf]
Matthew Howard, Stefan Klanke, Michael Gienger, Christian Goerick and Sethu Vijayakumar. A Novel Method for Learning Policies from Variable Constraint Data. Autonomous Robots, vol. 27, pp. 105-121 (2009). [pdf]
Chris Towell, Matthew Howard and Sethu Vijayakumar. Learning Nullspace Policies. Proc. IEEE/RSJ Intl. Conf. on Intelligent Robots and Systems (IROS 2010), Taiwan (2010). [pdf] http://youtu.be/dKfKK129tpk
Matthew Howard, Stefan Klanke, Michael Gienger, Christian Goerick and Sethu Vijayakumar. A Novel Method for Learning Policies from Constrained Motion. Proc. IEEE International Conference on Robotics and Automation (ICRA '09), Kobe, Japan (2009).[pdf]
Hsiu-Chin Lin, Matthew Howard and Sethu Vijayakumar, Learning Null Space Projections, Proc. IEEE International Conference on Robotics and Automation (ICRA 2015), Seattle,WA, USA (2015). [pdf]
Behaviour Generation in Humanoids
We have developed a method for learning potential based policies from constrained motion data. In contrast to previous approaches to direct policy learning, our method can combine observations from a variety of contexts where different constraints are in force, to learn the underlying unconstrained policy in form of its potential function.
Matthew Howard, Stefan Klanke, Michael Gienger, Christian Goerick and Sethu Vijayakumar. Behaviour Generation in Humanoids by Learning Potential-based Policies from Constrained Motion. Applied Bionics and Biomechanics, Vol. 5, No. 4, pp.195-211, Taylor and Francis (2008). [pdf]
Matthew Howard, Stefan Klanke, Michael Gienger, Christian Goerick and Sethu Vijayakumar. Learning Potential-based Policies from Constrained Motion. Proc. 8th IEEE-RAS International Conference on Humanoid Robots (Humanoids), Dejong, Korea (2008) [pdf]
TEDx Talk (2015): The robots are ready, are you?
Our Changing World Lecture: Shared Autonomy (2015)