Automatic Generation of Biped locomotion using Genetic Programming

Generating biped locomotion in robotic platforms is hard due to the complexity of the
the synchronization of several joints, while monitoring stability. Further, it is also expected to deal with the great
heterogeneity of existing platforms. The generation of adaptable locomotion further increases the complexity of the task.
Genetic Programming (GP) is used as an automatic search method for motion primitives of a biped robot, that
optimizes a given criterion. It does so by exploring and exploiting the capabilities and particularities of the platform.

In order to increase the adaptability of the achieved solutions, feedback pathways were directly included into the
evolutionary process through sensory inputs.

Results indicated that the utilization of feedback yielded slightly more stable locomotion, and that it provided the robot with the ability to deal with situations that otherwise could not. Controller solutions provided with feedback in the GP evolutionary process were able to climb and descend smooth slopes with a maximum inclination of 9.8 degrees. Solutions evolved without sensory information could not perform such task, but only to climb or descend the slope.

Resulting locomotion under in different scenarios are shown in the video attachments.

People involved in this project: 
Project status: 
Project in progress
bestcontroller1_down.avi1.06 MB
bestcontroller1_flat.avi894.04 KB
bestcontroller1_staged_updown.avi1.69 MB
bestcontroller1_up.avi1.05 MB
bestcontroller2_down.avi1.05 MB
bestcontroller2_flat.avi921.83 KB
bestcontroller2_fullslope.avi2.02 MB
bestcontroller2_plateau.avi2.03 MB
bestcontroller2_staged_updown.avi1.99 MB
bestcontroller2_up.avi1.06 MB
bestseed_flat.avi872.76 KB