Contribution to the Scheduling and Task Execution Control of an Autonomous Vehicle

Abstract

This dissertation offers a modest yet meaningful contribution toward ad- vancing control strategies for Electric Autonomous Vehicles (EAVs), through a comprehensive framework that integrates classical control structures, intel- ligent optimization, and adaptive control mechanisms. Initially, a comparative study of longitudinal speed profiles is conducted, for- mulating and evaluating S-curve and Sinusoidal trajectories to meet road safety and passenger comfort standards. The classical Proportional Integral Derivative (PID) controller is implemented for longitudinal speed tracking, with its performance subsequently enhanced through offline tuning via two nature-inspired Metaheuristic algorithms, Genetic Algorithm (GA) and the Jellyfish Search (JS) optimizer. Results demonstrated the superior optimiza- tion capabilities of Metaheuristics, with the JS optimizer exhibiting higher efficiency. To address the limitations of fixed-gain controllers in dynamic environments, Model Reference Adaptive Control (MRAC) strategies based on Lyapunov stability theory and Popov’s Hyperstability criterion are developed and eval- uated under scenarios involving parametric uncertainties and external dis- turbances. The Lyapunov-based MRAC consistently delivered superior ro- bustness. Extensive simulations within a MATLAB/Simulink environment were con- ducted to validate the proposed control strategies. The results highlighted the performance of each approach, classical PID control, optimized PID con- trol via Metaheuristic algorithms, and adaptive control across different test scenarios. The comparative analysis clearly identified the conditions under which each controller excelled. These findings provide valuable insights for the deployment of suitable control methods in Electric Autonomous Vehicle systems, contributing to the advancement of adaptive and intelligent driving technologies.