Research On Predictive Cruise Control System With Digital Map

The eco-powertrain control system leveraging the digital map is not only an important way for our country to realize the zero carbon goals,but also a necessary way for the traditional automobile industry to meet the carbon neutrality agenda without buying environmental credits.A traditional vehicle cruise system can track the set speed adaptively,thereby reducing the driver's driving fatigue.The emergence of digital maps endorse vehicles to predict the road conditions ahead and take actions in advance,thus giving more fuel-saving possibilities to the predictive cruise control system.However,many key scientific issues behind the cruise control integrated with digital maps remain unresolved.In particular,the theoretical method about throughout integration of digital maps and vehicle power systems is far from mature,which seriously affects the actual effect of predictive cruise control systems in mass-produced vehicles.On the one hand,it is difficult to integrate the gradient disturbance and speed limit constraint in the spatial domain of the digital map with the vehicle dynamics model in the time domain.Moreover,it is difficult to use the fuel cut-off mechanism in the vehicle powertrain system for predictive cruise control.On the other hand,the optimization problem considering the digital map information is complex to solve,and the quality of the optimal control solution on the limited computing power controller cannot be guaranteed.With this question,I went to the RACElab(Real-time Adaptive Control Engineering Lab)at University of Michigan,Ann Arbor to study predictive cruise control system with digital map via a two-year program funded by the China Scholarship Council.This paper aims to improve the energy efficiency of passenger vehicles,focusing at the challenges brought by the integration of digital map to the predictive cruise control system.The reaearch follows the roadmap of "vehicle system modeling-optimal control problem formulation – solving-real vehicle experimental verification".The contributions of this dissertation are as follows:First of all,for the predictive cruise control system aiming at energy saving,a digital map pre-analysis method is proposed,which can reduces the computational load of the controller.The challenges brought by the introduction of digital maps to the design of predictive cruise control algorithms are introduced.Considering the spatial sampling characteristics of the digital map,a vehicle power system model in the spatial domain is formulated.The model takes into account the energy-saving characteristics of the transmission system such as engine start-stop and fuel cut-off.Then,the vehicle model is calibrated using real vehicle data and bench test data.It will lay the foundation for the construction of the predictive cruise control system.Secondly,considering the gradient disturbance and speed limit constraints brought by the digital map to the predictive cruise control problem,a multi-objective predictive cruise algorithm based on the Utopia point tracking algorithm is proposed.The cost function is constructed to find the Pareto solution of the closest distance to the Utopia point on the Pareto frontier,which solves the problems of heavy calibration workload and poor robustness of fixed weights in the traditional weighted sum method.It can not only give a dynamic balance of fuel consumption and travel time in the dynamic environment and the improvement of comprehensive performance,but also provide an ideal solution for the design of the cost function of the multi-objective predictive cruise control algorithm under complex working conditions.In order to further improve the fuel-saving potential of the predictive cruise system,a predictive cruise optimization control algorithm incorporating the coasting mechanism is designed.Two different coasting strategies(fuel cut-off and engine start/stop)are proposed to reveal the potential benefit of eco-coasting using the road grade preview.Engine drag torque and energy cost used for engine restart are considered in the modeling to give a fair evaluation of the performance.The offline performance of these two coasting methods is evaluated through dynamic programming(DP)under various driving scenarios with different slope profiles.Offline simulation shows that the engine start/stop method outperforms the fuel cut-off method in terms of fuel consumption and travel time by getting rid of the engine drag torque.Then,online performance of these two coasting methods is evaluated using Mixed Integer Model Predictive Control(MIMPC).A novel operational constraint on the minimum off steps is added in the MIMPC formulation to avoid frequent switch of the integer variables which represent the fuel cut-off and the engine start/stop mechanism.Further,for the calculation of mixed integer nonlinear programming problems in the predictive cruis control system,a solving strategy composed of branch-and-bound method and sequential quadratic programming is proposed to solve the MINLP with ordinary differential equation constraints.Early branch pruning in the branch-and-bound method is implemented through the infeasible detection of the primal-dual problem.Code optimization is performed by using symbolic derivation calculation,thereby reducing the number of optimization operations required for each iteration.Based on the BBSQP solver,several acceleration methods are proposed to improve calculate efficiency by integrating the physical knowledge and controller's sequential characteristics.More specifically,based on the sequential nature of model predictive control,a warm-start strategy for branch-andbound is formulated.The minimal activating constraint of engine fuel cut-off is integrated into the tree search process to reduce the “useless” branch.The proposed eco-coasting strategy can realize real-time optimization of engine torque,braking force and engine fuel cut-off simultaneously.Finally,a vehicle experimental platform was built,and the fuel-saving effect of the predictive cruise control algorithm integrated with digital maps was verified through real vehicle experiments in real traffic environments.It is proved that the predictive cruise control algorithm proposed in this paper has high engineering application value.Code is available in Git Hub(https://github.com/yoya18).