Design of nonsmooth Kalman filter for compound sandwich systems with backlash and dead zone

In this article, a new nonsmooth Kalman filtering (KF) method is proposed for noise suppression of the compound sandwich with backlash and dead zone. Based on the characteristics of the system, a nonsmooth stochastic state-space equation is constructed. Then, nonsmooth KF is developed to adapt the nonsmooth stochastic state space model. The filter can switch automatically among different operating zones according to the operating conditions. Moreover, the convergence of the switchable compound nonsmooth KF is discussed. Subsequently, a motor-driving mechanical transmission system which can be described as a compound nonsmooth nonlinear sandwich system with backlash and dead zone is studied as an application case. Finally, the comparison between the proposed filtering scheme and conventional filter are presented. It is demonstrated that the nonsmooth KF has a better performance than the conventional one in terms of convergence and accuracy in state estimation under noisy environment.