Optimizing the Steering of Driverless Personal Mobility Pods with a Novel Differential Harris Hawks Optimization...

Area of Science:

• Robotics and Control Systems
• Artificial Intelligence
• Optimization Algorithms

Background:

• Steering performance is critical for automated driving systems (ADS), particularly in localization and path planning.
• Meta-heuristic optimization algorithms have shown promise in steering control applications.
• The Harris Hawks Optimization (HHO) algorithm is effective but has not been applied to steering control.

Purpose of the Study:

• To improve the steering performance of Ackermann personal mobility scooters.
• To introduce a novel meta-heuristic optimization algorithm, Differential Harris Hawks Optimization (DHHO).
• To model the steering encoder for enhanced control.

Main Methods:

• Supervised learning was used to model steering encoder sensor data from practical experiments.
• The DHHO algorithm was developed by incorporating mutation into the HHO exploration phase to enhance diversity.
• DHHO was implemented for real-time PID tuning to control scooter steering.

Main Results:

• DHHO demonstrated superior performance over HHO, PSO, BAS, and CMAES on CEC2021 benchmark functions.
• DHHO achieved significant error reduction compared to HHO in 10 and 20-dimension problems.
• Practical tests showed an 89.31% improvement in settling time for scooter steering, with no overshoot or steady-state error.

Conclusions:

• The DHHO algorithm offers a significant advancement in steering control for mobility scooters.
• DHHO provides a robust and efficient method for optimizing automated driving system performance.
• The proposed DHHO algorithm outperforms traditional control methods, demonstrating its practical applicability.