Simultaneous Estimation and Receding Horizon Sliding Control of Dog Clutch Coupling With Backlash for Rear Electrical Drivelines

Compared to friction clutches, dog clutches have the advantages of high efficiency, small size, low cost, and large transmission torque; therefore, they are selected for use in hybrid electric vehicles (HEVs) with a single electric motor (EM) driving the rear axle. In the dog clutch coupling process, the meshing of the teeth produces an impact, and dynamic control of the clutch faces problems of backlash and flexible shaft control. For an advanced dynamic study and subsequent improved performance, a novel dynamic estimation algorithm is proposed. This algorithm includes an observer for estimating the clutch torque, two piecewise-affine (PWA) estimators for estimating the torsion angle and half-shaft torque, and a state machine for indicating the backlash position. To demonstrate its feasibility, the convergence rate and observation error are investigated. Based on the information obtained through dynamic estimation, a second-order receding horizon sliding control (RHSC) approach is proposed. In dog clutch coupling, the speed deviation and tracing torque set points, which correspond to different backlash positions, are the control targets. Based on the second-order sliding mode structure, an optimization algorithm is developed and its convergence is studied. The performance of the estimator and control method is verified through road tests. The results show that the estimator can correctly indicate the system state and that the control method can improve the performance of the powertrain.