Autonomous race cars handle overtaking using a pipeline of sensing, prediction, planning, and control that is mathematically explicit and safety-constrained, whereas human drivers rely on learned heuristics, situational judgment, and often risk-tolerant instincts. Research by Chris Gerdes at Stanford University emphasizes the role of precise vehicle dynamics models and model predictive control to execute high-speed maneuvers while maintaining a formal safety envelope. Work by Raquel Urtasun at University of Toronto underscores how probabilistic perception and motion prediction reduce uncertainty about other agents before committing to an overtaking trajectory.
Sensing and prediction
Onboard sensor suites give autonomous racers a continuous, high-bandwidth view of the environment. Lidar, radar, and cameras feed algorithms that estimate position, velocity, and intent of nearby vehicles. Prediction modules produce probabilistic future trajectories for opponents, which planners use to evaluate overtaking opportunities. This contrasts with human drivers who infer intent from visual cues and experience and may tolerate higher ambiguity when deciding to pass. In marginal conditions such as spray, glare, or dust, the quality of prediction markedly changes the overtaking strategy.
Decision-making and control
Decision-making in autonomous systems is optimization-driven. Planners evaluate multiple candidate maneuvers against objective functions that balance lap time, collision risk, and regulatory constraints. Control systems execute selected trajectories using high-frequency feedback and vehicle models, enabling precise throttle, steering, and brake coordination during close passes. Humans rely more on feel, split-second judgment, and social signaling between drivers. The consequence is that autonomous overtakes tend to be more repeatable and auditable but may appear conservative compared with a skilled human pushing the limits.
Cultural and territorial factors matter. On closed circuits and in series designed for autonomy such as Roborace, rules and engineering focus encourage aggressive automated tactics and software-defined fairness. On mixed-use roads or in regions with strict liability laws, fleets emphasize conservative overtaking to prioritize safety and public acceptance. Environmental effects also differ: smoother, model-based maneuvers can reduce energy waste and tire wear, while aggressive human passes often increase mechanical stress.
These differences suggest a hybrid future where learned tactical behaviors inform planners, yet formal safety guarantees remain central. The combined insights of vehicle dynamics experts and perception researchers at institutions such as Stanford University and University of Toronto chart a path toward overtaking that is both fast and verifiably safe.