Hybrid vehicle transmission control using driver statistics

Abstract

A multi-layer control mechanism for optimizing performance metrics of a hybrid electric vehicle (e.g., fuel efficiency, drivability, NVH). A first layer generates a policy that defines target engine & motor operating settings for each of a plurality of possible driver demand inputs based on a predicted driver demand profile for a long-horizon period of time. A second layer determines a predicted 'short-horizon' driver demand—based, for example, on historical driver data and one or more environmental sensor inputs—and applies a corrective pre-adjustment to the operating settings of the vehicle in response to determining that a pre-adjustment is required in order to apply the target operating settings for the predicted driver demand. A third layer determines constraints to the operating settings required to comply with the additional performance parameters and limits the operating settings applied to the engine and motor(s) to feasible operating settings defined by the constraints.