Haptic Interfaces for Compensating Dynamics of Rescue Walking Robots

This paper presents a Neural Network approach to compensate dynamic terms, friction force in particular, of a rescue walking robot used in haptic interfaces. The impedance control through dynamic compensation of the friction force is studied, followed by the implementation of neural intelligent networks in the feed-forward loop in order to eliminate the corresponding terms in the dynamics, friction force in particular. The friction force model is analyzed using a general compensation method after which a trained Multi-Layer Neural Network is introduced in order to obtain an accurate friction model so that the movement of the walking robot feels free and unconstraint.