From Robots with Self-Inhibiting Modules to Habile Robots

Abstract

I describe a new robot control architecture based on self-organization of self-inhibiting modules. This architecture can generate a complex behavior repertoire. The repertoire can be performance-enhanced or increased by modular polyfunctionality and/or by the simple addition of new modules. I postulate that this architecture can evolve, in the hands of the designer, to a Habile robot, a version of Nilsson’s habile system. Each program module controls a joint motor or a motor for a pair of wheels. Every module estimates also the distance from a sensor (placed in the hand of the arm) to a beacon. If the distance is shorter or longer than a previously estimated distance, the module drives its motor in the corrective direction; if the movement produces no significant change in distance, the module self-inhibits. A self-organization emerges: once a module self-inhibits, any module can be the next to take control of the motor activity. The overall behavior of the robot corresponds to a reaching attention behavior. This is easily switched to an‘adverse’ attention behavior by changing the sign of the same parameter in each module. The addition of a “sensor-gain attenuation reflex” module and a “light orientation reflex”module provides an increase in the behavioral attention repertoire and performance enhancement. The ‘brain’ is actually providing action induction rather than action selection.