| Mission planning is the upper management module of the automatic driving system,which needs to plan the optimal global path and provide the relevant information of each mission quickly and autonomously with consideration of the user mission,dynamic traffic information and vehicle status.Pure electric vehicle has made great progress in the field of automatic driving because of its environmental friendliness and excellent control characteristics,and it is of great significance to properly estimate and optimize the energy consumption for its use and promotion.Aiming at the mission planning requirement of the autonomous electric vehicle,this paper is mainly focuses on the global path planning and the expected speed planning in order to optimize the travel time and energy consumption.First of all,the energy consumption model is established with the speed as independent variable based on the vehicle driving resistance formula,and the unknown parameter is identified by the recursive least square method with forgetting factor.Then,the time series neural network is used to estimate the battery SOC,and the capacity of the mission point can be predicted by combining the energy consumption model and the SOC estimation.The validity and accuracy of the algorithm are verified based on the data of real working condition.Secondly,considering the time dependence of traffic speed,the historical data and dynamic data are weighted by using the Sigmoid function to improve the reliability of the speed prediction.The Dijkstra algorithm is used to calculate the time and energy consumption weight matrices,so that the driving cost between any two road segments can be obtained.Then,the mission point sequence is optimized based on the dynamic programming algorithm,and the local searching strategy is designed to avoid the shortage of power.Also,A* algorithm is applied to realize the global path planning between adjacent mission points.The proposed methods are validated via sufficient simulations.Next,the path breakpoint model is introduced to calculate the road travel time and energy consumption,and the energy consumption model is simplified and analyzed based on the ratio of energy to mileage.The optimization objectives and constraints of speed planning are designed according to the mission requirements and traffic flow characteristics,and a non-dominated sorting genetic algorithm with elite strategy is proposed to solve the multi-objective optimization problem.Then,the objective function is normalized and the design principle of the weighting coefficient is given,so that the optimal solution is obtained by weighting multiple objective functions.The simulation results verify the stability and consistency of the algorithm.Finally,the mission planning function is tested based on the small scale road network and the non-parametric estimation is used to obtain the probability distribution of estimation error.An estimated value correction strategy based on the mean of error is proposed,which can not only improve the accuracy and reliability of the estimation information but also guarantee the effectiveness and practicability. |
