Design of computer experiments for open-loop control and robustness analysis of clutch-to-clutch shifts in automatic transmissions

The design of computer experiments is used to design an open-loop control for a clutch-to-clutch shift automatic transmission and to study the robustness of that control. The open-loop control design problem can be posed as an optimal control problem but because of the computational cost associated with each simulation and the complexity of the transmission model, classical results from optimal control theory are not a practically feasible approach for this problem. Likewise, due to the nonlinear, finite-time nature of this problem, existing approaches to robust analysis are not applicable. We apply the methods of design of experiments, surrogate modeling, and randomized algorithms for optimization to these problems and present some promising results