Adapting Quadruped Locomotion by Real Time Learning of Detection and Avoidance of an Obstacle

There is an increasing interest in conceiving robotic systems that are able to move and act in an un- structured and not predefined environment, for which autonomy and adaptability are crucial features. In nature, animals can be regarded as autonomous biological systems, which often serve as inspiration models, not only for their physical and mechanical properties, but also their control structures that enable adaptability and autonomy - for which learning is (at least) partially responsible.
This work proposes a system which seeks to enable a quadruped robot to learn to in real time detect and to avoid an obstacle in its path. The detection relies in a Forward Internal Model that estimates the robot’s perceptive information. The system adapts the locomotion in order to place the robot optimally before attempting to step over the obstacle, avoiding any collisions. Locomotion adaptation is achieved by changing control parameters of a Central Pattern Generator (CPG) based locomotion controller. The mechanism learns the necessary alterations to the stride length in order to adapt the locomotion by changing the required CPG parameter.
Both learning tasks occur in real time and together, define a Sensorimotor Map, which enables the robot to learn to step over the obstacle in its path. Simulation results show the feasibility of the proposed approach.

Attachment file presents a video where an example of the learning process is shown.

1 - The robot attempts to step over the obstacle too close to it and fails.
2 - In this second attempt, the robot reduces the stride length in order to avoid the previous situation. However the change causes the robot to attempt to step over the obstacle too far away, causing it to fail by colliding with the back side of the paw.
3 - In the last attempt, the adaptation of the stride length taking into consideration the previous experience, enabled the robot to attempt the task at the ideal distance, and thus clear the obstacle without any collision.

AttachmentSize
quadstepover2013_submitted.mpg8.1 MB