Dynamic modeling, clutch parameter adaptation, and control of automatic transmission for tracked vehicles

In this study, a novel TCU control structure and algorithms including also dynamic model of the transmission will perform the gear shifting of automatic transmissions used in tracked military vehicles are designed. In case of inherent wear of the clutches, the clutch parameters are estimated by means of an adaptation algorithm that uses only the information from the transmission input and output speed sensors. These estimated clutch parameters are used in clutch torque calculation is utilized by control algorithms. For the clutch speed control in the inertia phase, a PID control structure with a negative feedforward term that cancels the nonlinearities of the system is proposed. The feedforward term utilizes the estimated wear clutch coefficients in the calculation of friction coefficients. The parameters of the PID controller are optimized by a genetic algorithm. The designed all structure are firstly tested on a MATLAB/Simulink power train simulation model, and then tested using a unique transmission on a transmission dynamometer test bench to examine its suitability for real transmission scenarios. The performances of proposed PID control and a backstepping nonlinear controller that we constructed for the tracked vehicles are compared with and without clutch wear. Maximum tracking error is decreased by 98% with proposed adaptive PID control structure with nonlinear feedforward term compare to backstepping controller. In the case of clutch wear, the back-stepping control exists in the literature has been found to result in oscillation. By updating the estimated friction coefficient from adaptation algorithm, in the backstepping controller, it is concluded that these oscillations can be eliminated. Additionally, on-off solenoid valves, are widely used, are known to cause torque interruptions and jerks during gear shifting. It has been observed the control structure presented in this study reduced the jerks by one quarter.