Study On The Methods Of Vehicle Motion Planning And Powertrain System Control For Improved Fuel Economy

Energy saving and environmental protection are the global advocacy and practice goals.For automobile industry,the connected and energy-efficient vehicles are important ways to achieve the goal of ”low carbon,information,and intelligence”.Under this background,the methods of vehicle motion(route and speed)planning and powertrain system control was studied.This research was divided into two stages: stage 1 focused on the co-optimization of the vehicle route and speed,where the vehicle powertrain dynamics were regarded as the given parameter;stage 2 was the co-optimization of vehicle route,speed,and powertrain system control,where the optimization of powertrain system was studied with the vehicle motion planning method.First,a co-simulation model of traffic-vehicle dynamics(TVD)was established for the connected vehicles.The traffic model was built under Simulation of Urban MObility(SUMO)based on the actual traffic network;the powertrain models of traditional ICE vehicles and hybrid electric vehicles were built using Matlab/Simulink;the co-simulation models were generated based on these two kind models.The TVD models supplied the required data and verification platform for the following research.Second,a novel co-optimization problem of vehicle route and speed was studied,and a vehicle macroscopic motion planning(VMMP)method was proposed.The mathematical model of the VMMP problem was formulated using the traffic data and vehicle characteristics.This VMMP method tried to optimize the vehicle route and speed simultaneously to minimize the fuel consumption for the given origin-destination within an expected trip time.The genetic algorithm(GA)based co-optimization method was proposed to solve the economic route and speed.In addition,an adaptive real-time optimization strategy,along with penalty model of traffic lights was added to the VMMP method for the practice implementation.Four simulation studies were designed based on the TVD model and performed for ideal traffic environment,traffic light and jam situations,and different vehicle platforms,respectively.The simulation results showed that the proposed VMMP method was able to improve the vehicle fuel economy significantly.Comparing with the fastest route,the fuel consumption using the proposed VMMP method was decreased by up to 15%.Third,the variation characteristic and prediction method of traffic flow were studied,and a spatial-temporal traffic speed prediction model was proposed.According to the periodicity and spatial-temporal correlation of the traffic flow,a periodic average model and a spatial-temporal recursive space-time autoregressive one were proposed.And then,the spatial-temporal traffic speed prediction model was presented.Note that a recursive least squares method was put forward to identify the dynamic model parameters in real-time.Other static ones were selected by their prediction errors using the separation of variables.Combining with the VMMP method,a improved VMMP method was proposed based on the predicted traffic speed.This proposed method was verified by the TVD model using the actual traffic data from the Pe MS(Performance Measurement System)database.Forth,the co-optimization method of vehicle route,speed,and powertrain control was studied.A eco-driving method based on the VMMP and three-parameter gear-shifting schedule was proposed.For traditional ICE vehicles,the automatic shifting control is the main way to improve the fuel economy,since the acceleration and brake pedals are controlled by the driver.A three-parameter gear-shifting schedule based on vehicle throttle,speed,and road resistance coefficient was proposed.The road resistance coefficient,was a compound parameter consisting of the road grade and the rolling resistance coefficient.It was estimated in real time by the proposed multi-step recursive least-squares method.The dynamic programming and the moving least-squares method were adopted to optimize the gear sequences and generate the three-parameter gear-shifting schedule based on the standard driving cycles and vehicle actual driving cycle.The eco-driving method was designed by combining the VMMP method and three-parameter gear-shifting schedule.The proposed method was verified using the TVD model and on-road experiments were carried out using a heavy-duty vehicle with an AMT(automated manual transmission).The results showed that the three-parameter gearshifting schedule improved the fuel economy with a satisfactory acceleration performance.Last,an extended adaptive cruise control(ACC)system,called economic adaptive cruise control(EACC),was proposed to improve the fuel economy for HEVs by optimizing the vehicle route,speed,and powertrain control simultaneously.The HEV has two energy source and is more complicated.A hybrid powertrain controller combining feedforward and feedback modules was developed.The feedforward module was based on a proposed global power distribution strategy and the feedback module was based on a receding horizon linear quadratic tracking control.In addition,a mode switch based local optimization method using the minimum principle was added to modify the reference speed for passing,traffic jam and lights crossing.The co-simulation results based on the TVD model indicated that the proposed EACC was able to decrease the fuel consumption by up to 30% comparing with the fastest route using the power follower control strategy.