Search Results

Lane keeping assist (LKA) is an autonomous driving technique that enables vehicles to travel along a desired line of lanes by adjusting the front steering angle. Reinforcement learning (RL) is one kind of machine learning. Agents or machines are not told how to act but instead learn from interaction with the environment. It also frees us from coding complex policies manually. But it has not yet been successfully applied to autonomous driving. Two control strategies using different deep reinforcement learning (DRL) algorithms have been proposed and used in the lane keeping assist scenario in this paper. Deep Q-network (DQN) algorithm with discrete action space and deep deterministic policy gradient (DDPG) algorithm with continuous action space have been implemented, respectively. Based on MATLAB/Simulink, deep neural networks representing the control policy are designed. The environment as well as the vehicle dynamics are also modelled in Simulink.

As a global optimization method, dynamic programming (DP) can be employed to seek the optimal velocity with minimum energy consumption for EV on given driving cycles. Due to its terrible computational burden, conventional DP is not suitable for real-time implementation especially with higher dimensions. In this paper, we propose an iterative dynamic programming (IDP) approach to reduce computing time firstly. The IDP can obtain the optimal control laws alike the conventional DP by converging the optimal control strategy iteratively and save considerable computing time. Second, the developed IDP and model predictive control (MPC) are combined to establish a real-time cruising controller called IDP-MPC for an EV with terrain preview. In the predictive controller, we use the IDP to solve a constrained finite horizon nonlinear optimization problem.

Vehicle failure prediction technology is an important part of PHM (Prognostic and Health Management) technology, which is of great significance to the safety of vehicles and to improve driving safety. Based on the vehicle operating data collected by the on-board terminal (T-box) of the telematics system, the research on the state of vehicle failure is conducted. First, this paper conducts statistical analysis on vehicle historical fault data. Preprocessing procedures such as cleaning, integration, and protocol are performed to group the data set. Then, three indexes including recency (R) frequency (F), and days (D) are selected to construct a vehicle security status subdivision system, and K -Means algorithm is utilized to divide different vehicle categories from the perspective of vehicle value. Labeled information of vehicles in different security status are further established.