Robust BELBIC-Extension for Trajectory Tracking Control

摘要

In real-life trajectory tracking applications of robotic manipulators uncertain robot dynamics, external disturbances and switching constraints which cannot be accommodated for by a conventional controller affect the system performance. We suggested an additional control element combining sliding mode and bio-mimetic, neurologically-inspired BELBIC (brain emotional learning-based intelligent control). The former is invariant to internal and external uncertainties and guarantees robust behavior. The latter is based on an interplay of inputs relating to environmental information through error-signals of position and sliding surfaces and of emotional signals regulating the learning rate and adapting the future behaviour based on prior experiences and with the goal to maximize a reward function. We proofed the stability and the performance of the suggested control scheme through Lyapunov theory and numerical simulations, respectively.