Lessons learned from using some bio-inspired optimizers for real-time controller design for a low-cost electrohydraulic system

Display Omitted Controller design for an electrohydraulic system with proportional valve.Fuzzy control with suitable feedforward and bias compensations.Controller parameter optimization by Artificial Bee Colony technique.Real-time position demands of Regulation, tracking and composite types.Superior controller performance than that of other existing optimized servosystem. Well-designed searching procedures following natural processes have been developed for finding optimized solutions of complex systems. Here, a comparison of performances of some optimizers, namely differential evolution, genetic algorithm, bacterial foraging and artificial bee colony technique, have been carried out for designing a fuzzy-feedforward real-time controller of an electrohydraulic motion actuation system. The first two optimizers execute dominatingly exploratory search, while the latter two execute a combination of exploratory search with intensified exploitive search in prospective regions, thus providing faster convergence. The optimized controller has been designed by minimizing a response error integral for some standard displacement demands of the highly nonlinear system. The strong nonlinearities in the system arise from the friction of low-cost industry-grade cylinder and large deadband of rugged proportional valve. The convergences to the minimum of zero for a number of nonlinear functions have also been demonstrated for all the optimization processes. These optimizers with faster convergence rate have been shown to be robust against arbitrary demands like variable frequency sinusoidal demands and sinusoidal demands with superimposed log concave-convex variations.