面向无人驾驶车辆节能的迭代优化预测控制方法

An iterative optimization-based predictive control method for eco-driving of unmanned vehicles

该文提出一种数据驱动的无人驾驶重型车辆节能驾驶迭代优化预测控制(IOBPC)方法。基于历史数据构造并迭代更新终端状态约束集和终端油耗函数,并对约束集和终端油耗函数进行近似处理,提高算法计算效率;通过学习车辆状态轨迹和燃油消耗的关联机理,保证优化算法代价函数在迭代过程中单调递减并最终收敛,从而实现车辆燃油经济性的显著提升。结果表明:迭代优化预测控制器在多次迭代后使车辆轨迹收敛,燃油消耗减少了约10.2%,相比基于动态规划(DP)的节能驾驶策略,节能效果得到了进一步的提升,且调节参数较少,支持实时求解,更利于实际应用。

A data-driven Iterative Optimization-Based Predictive Control(IOBPC) method for energy saving of unmanned heavy vehicles was proposed. Based on the historical data, the terminal state constraint set and terminal cost function were constructed and updated iteratively. And the approximate treatment of the constraint set and terminal cost function improved the computational efficiency of the algorithm. By learning the correlation between vehicle state trajectory and fuel consumption, the cost function of the optimization algorithm was guaranteed to decrease monotonically and converge in the iterative process, so as to realize the significant improvement of vehicle fuel economy. The results show that the iterative optimization predictive controller makes the vehicle trajectory converged and reduces fuel consumption by about 10.2% after several iterations.Compared with the energy-saving driving strategy based on dynamic programming(DP), the energy-saving effect is further improved. Moreover, it has fewer adjustment parameters and supports real-time solution, which is more conducive to practical application.