Robust near time-optimal control

A robust control algorithm for a realization of nearly time-optimal control of double integrator plants is presented. The technique involves the combination of traditional bang-bang time-optimal control with the methods of sliding-mode control. The result is a nonlinear feedback scheme for maintaining a desired functional relationship among dynamic variables, in this case imitating the idealized dynamics of time-optimal control. The approach is nearly time optimal rather than exactly time optimal, since the bang-bang control components are restricted to values below full actuator saturation, thus reserving some actuator effort for compensation of disturbances and model imperfections. The algorithm blends smoothly into a linear controller near a stable goal location.< >