Could AI emulate a human driver in Formula One?

Since 2021, the Automotive Charity Sub-Committee, led by Assistant Giles Taylor, has established strong reciprocal relations with the Department of Mechanical and Aerospace Engineering at Brunel University. Each academic year, Liverymen of the Automotive Sub-Committee along with departmental head, Professor Diane Mynors, award an outstanding second year student with the Coachmakers’ Advanced Automotive Engineering Research Bursary. This award enables the wining student to carry out an on-campus research project during their summer vacation period.

This year, we are delighted to report that BEng Mechanical Engineering student Benjamin Samuels used the award to investigate the future impact of AI on Formula One racing. Using MATLAB code, he was able to compare human inputs to that of an AI driver during a given race course simulation. He is invited to receive his award from the Master Coachmaker at the Automotive Dinner next March. Here he tells us about his research.

Advanced Automotive Engineering Research Bursary 2023

My journey into the study of mechanical engineering was driven by a deep fascination with physics and maths and how these tools can be used to innovate our society, writes Benjamin. This fascination transformed into a deep passion for all things engineering and problem solving through implementing logic and understanding challenges. As the pace of technological innovation increases, the field of artificial intelligence has exploded into relevancy. My ambition for this project was clear: to harness the precision and adaptability of artificial intelligence and meld it with the dynamism of motorsports. The use of AI is not just limited to racing but in exploring its implications across a multitude of dynamic systems.

In collaboration with the Brunel Autonomous Racing Team, our shared objective was to compete in the Formula Student AI competition. The centrepiece of our project was a vehicle control system that is driven by AI in conjunction with MATLAB. My personal goal was to discover if an AI system could possibly emulate or surpass a human driver under the same conditions of a Formula One race.

The research was conducted across various simulation environments, beginning with a straight track and culminating in the complex circuits of Formula One. Our approach entailed using data derived from MATLAB code to determine vehicle trajectories and reference points. After looking into a multitude of path planning algorithms and controllers the pure pursuit controller emerged as a suitable controller. This controller uses a look-ahead distance to guide the vehicle and ensure it follows the optimal path.

The use of this controller alongside the vehicle controls that we designed showed great promise in the potential of AI-driven systems. The system showcased adaptability and consistency across varied conditions. To evaluate the system’s effectiveness, we monitored metrics such as lap time, yaw rate, and velocity profile, then compared them to anticipated results. Also, visualisation from 3D simulations and 2D XY plots, provided a comprehensive understanding of the potential of artificial intelligence in the future of Formula One racing.

More information about our awards, scholarships and bursaries can be found here: https://www.coachmakers.co.uk/our-awards-and-bursaries/