Optimized PID Tuning in Longitudinal Control of Electric Autonomous Vehicles: A Comparative Study of Jellyfish Search and Genetic Algorithm

Tuning PID controllers to enhance longitudinal control and speed planning of Electric Autonomous Vehicles is a challenge, which can be effectively addressed by evolving metaheuristic algorithms. This paper evaluates the performance of the Jellyfish Search (JS) optimizer and Genetic Algorithms (GA) against conventional PID tuning methods in longitudinal control systems. A modified objective function combining Time-weighted Absolute Error (ITAE) and Integral Square Error (ISE) is proposed based on the weighted sum method, aiming to balance performance metrics and overcome the limitations of conventional objective functions. This function is optimized by both genetic and jellyfish techniques. The simulations are conducted through realistic scenarios in accordance with road safety standards, using a sinusoidal profile as speed reference. The results demonstrate the effectiveness of both GA and JS in outperforming conventional PID, achieving zero overshoot, reduced settling times, and lower steady-state errors. Observably, JS optimizer exhibits a slight advantage over GA in the overall performance, especially fast convergence. These outcomes validate the contribution of the proposed approaches in enhancing the field of autonomous driving.