OptLayer - practical constrained optimization for deep reinforcement learning in the real world

T.-H. Pham, G. De Magistris, R. Tachibana, IEEE International Conference on Robotics and Automation (ICRA), 2018 (to appear)

While deep reinforcement learning techniques have recently produced considerable achievements on many decision-making problems, their use in robotics has largely been limited to simulated worlds or restricted motions, since unconstrained trial-and-error interactions in the real world can have undesirable consequences for the robot or its environment. To overcome such limitations,... [Read More]
Tags: research

Hand-object contact force estimation from markerless visual tracking

T.-H. Pham, N. Kyriazis, A. A. Argyros, A. Kheddar, IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI)

We consider the problem of estimating realistic contact forces during manipulation, backed with ground-truth measurements, using vision alone. Interaction forces are usually measured by mounting force transducers onto the manipulated objects or the hands. Those are costly, cumbersome, and alter the objects’ physical properties and their perception by the human... [Read More]
Tags: research

Multi-contact interaction force sensing from whole-body motion capture

T.-H. Pham, S. Caron, A. Kheddar, IEEE Transactions on Industrial Informatics (TII), 2017

We present a novel technique that unobtrusively estimates interaction forces exerted by human participants in multi-contact interaction with rigid environments. Our method uses motion capture only, thus circumventing the need to setup cumbersome force transducers at all potential contacts between the human body and the environment. This problem is particularly... [Read More]
Tags: research

Whole-body contact force sensing from motion capture

T.-H. Pham, A. Bufort, S. Caron, A. Kheddar, "Whole-Body Contact Force Sensing From Motion Capture", in IEEE/SICE International Symposium on System Integration (SII 2016)

In this paper, we challenge the estimation of contact forces backed with ground-truth sensing in human whole-body interaction with the environment, from motion capture only. Our novel method makes it possible to get rid of cumbersome force sensors in monitoring multi-contact motion together with force data. This problem is very... [Read More]
Tags: research

Capturing and reproducing hand-object interactions through vision-based force sensing

T.-H. Pham, A. Kheddar, A. Qammaz, A. A. Argyros, IEEE ICCV Workshop on Object Understanding for Interaction (OUI 2015)

Capturing and reproducing hand-objects interactions would open considerable possibilities in computer vision, human-computer interfaces, robotics, animation and rehabilitation. Recently, we witnessed impressive vision-based hand tracking solutions that can potentially be used for such purposes. Yet, a challenging question is: to what extent can vision also capture haptic interactions? These induce... [Read More]
Tags: research