Reactive trotting with foot placement corrections through visual pattern classification

Agile robot locomotion on rough terrain is highly dependent on the ability to perceive the environment. In this paper, we show how the interaction between a reactive control framework and an online mapping system can significantly improve the trotting performance on irregular terrain. In particular, this new locomotion controller increases the stability of the robot and reduces frontal leg and shin collisions with obstacles by correcting in realtime the foothold locations. The mapping system uses an RGB-D sensor and a motion capture system to build a three dimensional map of the surroundings of the robot. While the robot is trotting, the control framework requests in advance a local heightmap around the next nominal foothold position. Then, an optimized foot placement location is estimated by applying visual pattern classification on the acquired heightmaps, and the leg endpoint trajectory is modified accordingly. The foothold correction is performed independently for each leg. To show the effectiveness of our approach the controller was tested both in simulation and experimentally with our 80 kg hydraulic quadruped robot, HyQ. The results show that visual based reaction through pattern classification is a promising approach to increase locomotion robustness over challenging terrain.

[1]  Larry H. Matthies,et al.  Robust multi-sensor, day/night 6-DOF pose estimation for a dynamic legged vehicle in GPS-denied environments , 2012, 2012 IEEE International Conference on Robotics and Automation.

[2]  Stefan Schaal,et al.  Learning, planning, and control for quadruped locomotion over challenging terrain , 2011, Int. J. Robotics Res..

[3]  Kurt Konolige,et al.  Calibrating a Multi-arm Multi-sensor Robot: A Bundle Adjustment Approach , 2010, ISER.

[4]  Paul Filitchkin,et al.  Feature-based terrain classification for LittleDog , 2012, 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems.

[5]  Darwin G. Caldwell,et al.  A reactive controller framework for quadrupedal locomotion on challenging terrain , 2013, 2013 IEEE International Conference on Robotics and Automation.

[6]  Ferdinando Cannella,et al.  Design of HyQ – a hydraulically and electrically actuated quadruped robot , 2011 .

[7]  Darwin G. Caldwell,et al.  Path planning with force-based foothold adaptation and virtual model control for torque controlled quadruped robots , 2014, 2014 IEEE International Conference on Robotics and Automation (ICRA).

[8]  Wei Wang,et al.  Obstacle crossing with stereo vision for a quadruped robot , 2012, 2012 IEEE International Conference on Mechatronics and Automation.

[9]  Douglas Hackett,et al.  An overview of the Defense Advanced Research Projects Agency’s Learning Locomotion program , 2011, Int. J. Robotics Res..

[10]  Darwin G. Caldwell,et al.  Vision enhanced reactive locomotion control for trotting on rough terrain , 2013, 2013 IEEE Conference on Technologies for Practical Robot Applications (TePRA).

[11]  Darwin G. Caldwell,et al.  LOCAL REFLEX GENERATION FOR OBSTACLE NEGOTIATION IN QUADRUPEDAL LOCOMOTION , 2013 .

[12]  Larry H. Matthies,et al.  High fidelity day/night stereo mapping with vegetation and negative obstacle detection for vision-in-the-loop walking , 2013, 2013 IEEE/RSJ International Conference on Intelligent Robots and Systems.

[13]  Darwin G. Caldwell,et al.  Onboard perception-based trotting and crawling with the Hydraulic Quadruped Robot (HyQ) , 2013, 2013 IEEE/RSJ International Conference on Intelligent Robots and Systems.

[14]  Heiko Hirschmüller,et al.  Stereo-vision-based navigation of a six-legged walking robot in unknown rough terrain , 2012, Int. J. Robotics Res..

[15]  Ashutosh Kumar Singh,et al.  The Elements of Statistical Learning: Data Mining, Inference, and Prediction , 2010 .

[16]  Alfred A. Rizzi,et al.  Autonomous navigation for BigDog , 2010, 2010 IEEE International Conference on Robotics and Automation.

[17]  Darwin G. Caldwell,et al.  Terrain mapping with a pan and tilt stereo camera for locomotion on a quadruped robot , 2014 .

[18]  Robert Tibshirani,et al.  The Elements of Statistical Learning: Data Mining, Inference, and Prediction, 2nd Edition , 2001, Springer Series in Statistics.

[19]  Darwin G. Caldwell,et al.  Quadruped robot trotting over irregular terrain assisted by stereo-vision , 2014, Intell. Serv. Robotics.

[20]  Andrew Y. Ng,et al.  Stereo vision and terrain modeling for quadruped robots , 2009, 2009 IEEE International Conference on Robotics and Automation.