A valve reference position-learning device for an internal combustion engine, which is capable of accurately learning the reference position of a valve both before and after the warm-up of the engine, while properly reflecting thereon a deviation of the reference position caused by the thermal elongation of the valve and drive system after the warm-up of the engine. The valve reference position-learning device calculates and updates a learned value of the reference position of the wastegate valve both at and after the start of the engine, by first and second learnings, respectively. The learned value calculated by the first learning is stored as an existing learned value. Further, at the start of the engine, if the calculation of the learned value is not completed, the learned value is set to the existing stored learned value.

A vehicle comprises an engine, a throttle body including a passage section and a throttle valve, and a throttle valve controller which controls opening or closing of the throttle valve, wherein the throttle valve controller includes: a memory section which contains therein a medium output control command such that the medium output control command continues to be stored in the memory section in a state in which electric power supply to the throttle valve controller is ceased, the medium output control command indicating an opening degree of the throttle valve, corresponding to a preset medium output range, and wherein the throttle valve controller controls opening or closing of the throttle valve, in accordance with the medium output control command read from the memory section, when an engine output meets a medium output condition in which the engine output is controlled to fall into the medium output range.

There is obtained a throttle learning control apparatus that raises the reliability of throttle fully closed and opened learning even when a self-shutoff delay abnormality occurs and can suppress various defects that occur when a throttle fully closed learning value or a throttle fully opened learning value is not updated. After a self-shutoff delay abnormality occurs, an electronic control unit implements fully closed learning 1 and fully opened learning 1 while the engine is operated in a non-engine-stall period and implements fully closed learning 2 and fully opened learning 2 for compensating the fully closed learning 1 and fully opened learning 1. Fully closed learning is completed when the fully closed learning 1 or the fully closed learning 2 has once been implemented; similarly, fully opened learning is completed when the fully opened learning 1 or the fully opened learning 2 has once been implemented.

Various methods for determining a holding current based on a pressure differential across a wastegate valve are provided. In one example, a method of operating a wastegate comprises determining a holding current with which to hold a wastegate valve at a desired position, the holding current determined based on a pressure differential across the wastegate valve.