Modeling and hybrid position-force control of walking modular robots

Abstract

The paper presents new concepts and approaches of applied mathematics in modeling and hybrid positionforce control for walking modular robots, developing an open architecture real time control multiprocessor system, in view of obtaining new capabilities for walking robots. A strategy is developed for the dynamic control of the movement of walking robots using the Zero Moment Point method by processing inertial information of force, torque and tilting and by implementing intelligent high level algorithms. For the modeling of some of the movements of the walking robot, the equation of a simple inverted pendulum was adopted, with a joint in the single support phase, which opposes the dampening forces of the leg joints. For other movements new control strategies were conceived through sequential analyses and local modeling of the movement parameters. The complexity of the movement mechanism of a walking robot was taken into account, being a repetitive tilting process with numerous instable movements and which can lead to its turnover on rough terrain. The control system architecture for the dynamic robot walking is presented in correlation with the control strategy.