A Robust Statistical Collision Detection Framework for Quadruped Robots

In order for it to achieve its tasks in an effective manner an autonomous mobile robot must be able to detect collisions and recover from them quickly. This paper proposes a new solution to the problem of detecting collisions during omnidirectional motion of a quadruped robot with an internal accelerometer considering it as an instance of general signal processing and statistical anomaly detection problem. We find that temporal accelerometer readings examined in frequency domain are good indicators of regularities (normal motion) and novel situations (collisions). In course of time, the robot builds a probabilistic model that captures its intrinsic properties while walking without obstruction and uses that model to determine whether there is an abnormality in case of an unfamiliar pattern. The approach does not depend on the walking algorithm used and walk characteristics, and is insensitive to the surface texture that the robot walks on as long as the surface is flat.