Hybrid adaptive impedance force controller using bang-bang and Particle Swarm Optimization approaches

Force control is one of the most challenging controls in robot manipulators. In this control scheme, the system dynamic does not only depend on actuator dynamic but also environment. Non-adaptive controller is not sufficient to efficiently regulate the plant when the environment such as manipulated object, contact point, etc. is changed. Adaptive controller is able to deal with this problem; however, its response in the learning (adaptation) period is often unsatisfactory. In some cases, this undesired response may damage the environment and actuator. To overcome this problem, our proposed technique applies Particle Swarm Optimization (PSO) to achieve the desired response. Hybrid structure is adopted to reduce the problem of unlearned response. The controller structure is based on the concept of impedance control which the controller regulates the system to act as the pre-specified impedance dynamics. Simulation results show that our proposed technique is applicable and superior to the conventional learning system.