This project is developping novel bio-inspired methods to allow complex, articulated robots to manage challenging environments, by relying on cheap, off-the-shelf sensors and computation. In doing so, this work will allow novel deployments of mobile robots, in particular for search-and-rescue scenarios, as well as for inspection or mapping of difficult to reach and/or dangerous areas where humans should not venture. This work is centered around the use of central pattern generators (CPGs), a bio-inspired, decentralized locomotive framework coordinating a legged robot’s shoulder joints as coupled oscillators. The locomotive patterns output by CPGs can then be adapted based on sensory feedback. In particular, this work will investigate the use of torque sensors, 2D/3D cameras, as well as LiDARs (or combinations thereof), onboard a series-elastic hexapod robot. By validating these controllers on hardware and under real-life conditions, the ultimate goal of this work will be to to push the current barriers and offer novel ways to deal with actual disaster scenarios or industrial inspection tasks.
Project reports from previous years:
Related prior projects: