Real-time Optimal Automotive Control For Intelligent Energy Conservation And Road Test

The development of automotive industry brings benefits to modern transportation meanwhile arouses energy shortage problem.The intelligentization of automotive basing on automation and information technologies brings new opportunities for energy-efficient control.Under this circumstance,the intelligent energy-efficient automotive control has drawn increasing attention.The core idea is that,by comprehensively evaluating the impact of road conditions on energy efficiency with the navigation,high-precision map and prediction of future traffic information,the driving decision and control output of powertrain can be optimized,and the matching relation between vehicle movement and road conditions,traffic conditions,as well as vehicle performance can be rationalized,which can greatly improve the energy efficiency.A mount of simulation works have shown that intelligent energy-efficient control can reduce at least 15-20% energy consumption with the optimization of driving strategy and powertrain by utilizing the intelligent traffic information from vehicle-to-vehicle(V2V)and vehicle-to-road(V2R)communication.However,there are still tremendous challenges existing in the process of promoting the real-life application of the intelligent energy-efficient automotive control,as follows: 1)How to fully utilize the obtained external traffic and geographic information combined with velocity profile optimization and powertrain control,and establish an unitized system architecture consisting of intelligent information and powertrain system for best energy saving,is a most urgent problem to be solved.2)The introduction of abundant external information the complicated driving environment,such as access restriction of traffic lights and nonlinear dynamic characteristics of powertrain system,has made the energy-efficient control much more complicated,including two aspects.On the one hand,different from traditional control systems,the energy-efficient control under intelligent driving environment has a conflict between the “fast” control of the powertrain system and the “long” prediction horizon in the driving strategy optimization.On the other hand,the primitive on-board hardware computing ability(hundreds of megabytes)poses a great challenge for the real-time control of the system.Therefore,how to realize the fast solution and real-time calculation of complex nonlinear optimal control problems for automotive energy-efficient control is another challenging difficulty.Focusing on the challenging problems in intelligent automotive energy-efficient control under the environment of smart city and the intelligent transportation system,the thesis will conduct systematic researches from theoretical methods to engineering practice on two main perspectives,that the effective usage of intelligent road information,and the formulation and real-time control of energy-efficient system.And finally,experiment platform basing on DongFeng AX7 is established to conduct road test and verify the theory with application research.Firstly,for the automotive energy-efficient control with large space-time preview span in the macro/micro traffic and long prediction time horizon,a fast solution algorithm for nonlinear model predictive control(MPC)with long prediction horizon is proposed,basing on the characteristics of powertrain system.The original nonlinear optimal control problem is transformed into two-point boundary value problem by introducing co-states.Then the explicit formulation of optimal control sequence under the initial values of co-states is obtained by using the characteristics of automotive powertrain system,to realize the real-time solution of embedded MPC controller.Secondly,considering the large variations of automotive subsystems in time and space scales,a hierarchical MPC framework with multiple space-time span is established,and hierarchical velocity profile optimization and energy management strategy are proposed,which integrate multi-source heterogeneous traffic information into different system levels and reduce the complexity of multi-objective multivariate optimal control problem and improve the computational efficiency.Then,for energy conservation in the urban traffic environment,an eco-driving predictive cruise control system is proposed with consideration of peripheral automotive dynamics and future road information.By predicting the future driving behavior of the preceding vehicle,the dynamic security constraints are established with road slope,curvature and speed limits considered in the optimization problem,to realize the dynamical optimization of engine torque,brake force and gear position under varied driving conditions.Finally,basing on the theoretical analysis and practical researches of real-time automotive intelligent energy-efficient control,a predictive cruise control system that utilizes the future road information is developed.On the basis of DongFeng AX7 experiment platform,road test with cumulative mileage over 3300 km is accomplished.The results show that 1)The system satisfy the real-time demands of vehicle-level hardware;2)Compared with the traditional adaptive cruise system,8-10% energy saving can be realized.Through these works,the systematic researches from theoretical analysis to engineering practice are accomplished,and the proposed methods have been clearly and effectively verified.Based on the current researches,the future work will focus on the fast solution algorithm for nonlinear MPC with path constraints,automotive energy-efficient task planning,and multi-vehicle coordination energy-efficient control.