Optimization Of The Vehicle Trajectory On The Hilly Road And The Intersection Based On Curvilinear Functions Of Eco-operating Modes

Hilly roads and intersections have been the point of high fuel consumptions because of special terrain conditions and traffic conditions.To change the driving behaviors of drivers and improve the fuel economy,the eco-driving concept has increasingly attracted attentions in light of the increased level of automation and intelligence of vehicles.However,most of existing studies on the eco-driving on the road segment have focused on the trajectory optimization on the flat road and few are applicable to the hilly road,due to the difficulty to obtain the road gradient data.Moreover,existing studies on the eco-driving at the intersection have focused on only optimizing the trajectory of the vehicle approaching the intersection,and few have taken into account the whole trajectories through the intersection inclusive of the vehicle approaching,passing through,and departing the intersection.Meanwhile,several artificial intelligent algorithms have been adopted to solve the eco-driving models,however,due to their complexities,they often exhibited low efficiency and low robustness.As a result,such an eco-driving trajectory optimization approach is not beneficial to practical applications as a real-time strategy.In this context,this dissertation presented the development of a novel approach to optimizing the trajectory of a vehicle running on a hilly road and at an intersection based on curvilinear functions of the eco-operating modes.By developing vehicle operating mode models for the eco-driving,the dissertation not only presented optimization schemes for generating an ecological trajectory to the vehicle running on the hilly road and the intersection with or without a queue,but also provided an application study on the special road segment composed of a hilly segment connected with an intersection.The dissertation includes the following contents:(1)Based on the latitude,longitude,and altitude data collected by the GPS receiver,a practical approach was developed to quantify the road gradient by dividing the GPS points into pools.As the latitude and the longitude data is inaccurate while the vehicle stops and the resolution of the altitude data is often an integer,the research designed a data correction method and a quality control process.Meanwhile,after analyzing the accuracy of the estimated road gradient using different lengths of the pool,it was shown that the estimated road gradient was closer to the real value when the pool length was set to 60 m.The average RMSE(Root-Mean-Square Error)was 0.0017.(2)For the vehicle trajectory is closely related to the acceleration,the characteristics of the acceleration were analyzed,including the range of the acceleration,the change of the acceleration with the speed,and the distribution pattern of the jerk,i.e.the variation rate of the acceleration with the time.Based on the statistical analysis,the acceleration was mainly distributed in[-2.15,1.72](m/s2),and the distribution probability decreased along with the increasing of the acceleration.Further,the distribution of the jerk was fitted well by the Laplace function,and the average R-square for testing the goodness of fitting was 0.975.(3)To optimize the trajectory of the vehicle running on a road segment,the curvilinear function of the eco-accelerating mode was first developed by analyzing the profiles optimized by the multiple population genetic algorithm.Then,the curvilinear function of the eco-braking mode was generated and proved analytically to be effective.Moreover,the curvilinear function of the cruise mode was developed after analyzing both the long-distance travel and the short-distance travel.Based on the above results,the trajectory optimization scheme was developed by using the curvilinear functions of the eco-operating modes.Meanwhile,it was shown that the generated trajectory based on the curvilinear function of the eco-accelerating mode was similar to the one generated by the multi-population genetic algorithm,thus the difference of the fuel consumptions was relatively small.(4)The proposed approach was adopted on a simulation platform to analyze its ability to reduce the fuel consumption under different traffic volumes and market penetration rates of Ecological Speed Drive(ESD)vehicles.It was shown that when the market penetration was fixed,the fuel savings decreased as the traffic volume increased.But when the traffic volume was fixed,more fuel savings were achieved with the increase of the penetration rate.In addition,the ESD vehicles had a positive effect on the Normal Speed Drive(NSD)vehicles on a single lane.But on multiple lanes,the effect was almost none.(5)To optimize the trajectory of the vehicle across the intersection,the research analyzed the blocking speed and the dissipating speed of vehicles' queuing at the intersection based on the LWR(Lighthill,Whiteham,and Richard)theory.Then,an approach was developed to calculate the queue length.Subsequently,approaches were developed to generate ecological trajectories under different traffic signal statuses based on the above curvilinear functions of eco-operating modes for the intersection both with a queue and without a queue.Finally,the fuel savings of a single vehicle as well as the traffic flow were analyzed based on the simulation platform VIS SIM.In the end,by combining the trajectory optimization methods for the hilly road and the intersection,the research selected a road segment near the Hangtian Bridge on the Fucheng Road in Beijing as a case study.The results showed that the average fuel saving rate were 5.9%and 16.1%for the intersection without a queue and with a queue.