Hierarchical speed control for autonomous electric vehicle through deep reinforcement learning and robust control

Abstract

For the speed control system of autonomous electric vehicle (AEV), challenge happens with how to determine an appropriate driving speed to satisfy the dynamic environment while resisting uncertainty and disturbance. Therefore, this paper proposes a robust optimal speed control approach based on hierarchical architecture for AEV through combining deep reinforcement learning (DRL) and robust control. In decision-making layer, a deep maximum entropy proximal policy optimization (DMEPPO) algorithm is presented to obtain an optimal speed via dynamic environment information, heuristic target entropy and adaptive entropy constraint. In motion control layer, to track the learned optimal speed while resisting uncertainty and disturbance, a robust speed controller is designed by the linear matrix inequality (LMI). Finally, simulation experiment results show that the proposed robust optimal speed control scheme based on hierarchical architecture for AEV is feasible and effective.

Optimal Model Reference Lead Compensator Design for Electric Vehicle Speed Control Using Zebra Optimization Technique

“…There have been many approaches proposed to analyze and control the speed of the electric vehicle system, including the design of the of PID controller with speed constraints [1],fuzzy PD-I controller based on the adaptive black hole optimization [10], fuzzy PID controller based on the Grey Wolf Optimization (GEO) and Particle Swarm Optimization (PSO) techniques [11],fractional order PID controller based on Extra-X second-generation Current Conveyor (EX-CCII) [2],integration and PD algorithm approach for optimizing shifting control in electric clutchless automatic mechanical transmissions [12], multi-mode fuzzy control strategy with real-time recognition of driving patterns for enhancing the fuel efficiency in the hybrid electric vehicles [13], fuzzy fractional order PID controller based on Ant Colony Optimization (ACO) [14], torque adaptive drive anti-skid control by addressing the lateral wind interference and improving the driving performance [15],type-2 fuzzy neural network [16],robust optimal control based on the deep reinforcement learning [17], and robust Anti-Lock Braking System (ABS) control method with real-time road recognition for enhanced performance [18].Furthermore, preceding research has extensively explored the construction of lead compensators through the utilization of optimization algorithms. This includes investigations into lead compensator design achieved through methods such as the application of genetic algorithms (GA) and particle swarm optimization (PSO) [7], the employment of the Most Valuable Player Algorithm (MVPA)for lead compensator synthesis [19], as well as the formulation of an adaptive fuzzy lead-lag controller utilizing a customized grasshopper optimization algorithm (MGOA) [20].…”

Online Gain Tuning Using Neural Networks: A Comparative Study

“…Many studies have been conducted with deep reinforcement learning in agriculture [11][12][13], but some do not compare with existing classic control, and they do not consider the possible approaches that hybridize both classic predictive control and deep reinforcement learning.…”