Real-Time Speed Trajectory Planning for Minimum Fuel Consumption of a Ground Vehicle

One of the most effective methods to reduce fuel consumption is modifying or shaping the driving loads of ground vehicles. For an automatic driving vehicle, the "pulse-and-glide" strategy is known to be optimal, and many analyses have been performed on this strategy in offline optimization. However, there is no pulse-and-glide algorithm that is implementable in real-time without loss of optimality. This paper presents the solution structure of an optimal control problem for the pulse-and-glide driving strategy. The solution structure is used in designing a real-time implementable algorithm that determines the optimal speed trajectory and the optimal gear shifting trajectory for pulse-and-glide operation. The optimal problem is formulated in a speed-acceleration domain (v-a domain) not in a time-velocity domain, which greatly reduces the complexity of the problem. Therefore, an optimal solution can be achieved by using the necessary conditions with a small number of grid search computations. The algorithm achieved 3%-5% fuel savings compared to the cruising case. In contrast, the previous real-time implementable pulse-and-glide method found in other literature achieved only 2%-3% savings with the same road conditions. By introducing a new domain for the problem formulation, the designed algorithm can be used in real-time without loss of optimality.