Online energy management for a solar car using pseudospectral methods for optimal control

Abstract

This paper deals with the problem of energy planning for a solar car race. Therein, an electric vehicle powered only by means of a battery pack and the sun's energy must cross usually long distances in several days, facing disturbances along the road such as road slopes, winds, cloudy weather, and vehicle failures. Because the battery pack energy is insufficient to end the race, teams must be very careful with their energy consumption planning, so as to drive as fast as possible to reach the goal in minimum time without emptying the battery pack. The proposed methodology consists of a real-time implementation that measures the current state of the vehicle and then executes a quick energy planning for the long term based on the available battery energy and current solar radiation forecast, afterwards executing a continuous optimal control problem stage for the short term solved by means of pseudospectral methods. This last inclusion allows for fast calculation of trajectory updates while keeping modeling detail.