Reactive locomotion of a hexapod for navigation across irregular ground

Abstract

In controlled environments, the hexapod limbs actuation can be controlled as a closed system. However, the increase of the terrain complexity implies an adaptation of their trajectory based on the robot interactions with the environment. Thus, the implementation of terrain data to the legs actuation potentially improves the hexapod quasi-static stability in these scenarios. This paper presents an adaptive control system based on the limbs reactive behavior for navigation across complex environments. Through force sensors placed on the foot-tips, the model detects the foot-ground interactions and adjusts the limbs trajectory accordingly. Furthermore, to ensure that the robot posture remains stable throughout locomotion, an impedance control is implemented in each limb. The proposed control architecture was tested in an irregular ground through dynamic simulations with five different configurations. Through result analysis, an optimized model was achieved which reduces the oscillations of the torso and slippage of the feet when walking across obstacles.