An Evaluation Method for the Intelligence Level of Gear Shift Decision Strategies in Automatic Transmission Vehicles Under Complex Scenarios
摘要
Gear shift decision strategies of automatic transmission vehicles significantly impact drivability, fuel economy, and smoothness. To evaluate their intelligence under complex conditions, this paper proposes a method based on natural driving data, integrating Analytic Hierarchy Process (AHP), entropy weight method, game theory, and fuzzy comprehensive evaluation. First, gear shift intelligence is defined, with primary indicators built on economy, generalized smoothness, power, and auxiliary braking. Second, using gaussian mixture model (GMM) and data mining, 40 secondary indicators are formulated from primary values to create a multi-dimensional evaluation system. Then, AHP and entropy weight method calculate subjective and objective weights, combined via game theory for comprehensive weights. Based on baseline strategies, membership functions are constructed, and fuzzy comprehensive evaluation forms the final method. Experiments confirm its effectiveness in assessing intelligence, guiding optimizations, and enhancing vehicle performance and intelligence.