Pure Pursuit Guidance for Car-Like Ground Vehicle Trajectory Tracking

Trajectory tracking guidance and control for nonholonomic (car-like) Autonomous Ground Vehicles (AGV), such as self-driving cars and car-like wheeled mobile robots, is a more challenging control problem than path following control, because the latter does not impose a speed requirement on the vehicle motion. The tracking error dynamics along the nominal path are nonlinear and time-varying in nature, which need to be exponentially stabilized. This paper presents a Line-of-Sight (LOS) Pure-Pursuit Guidance (PPG) trajectory design algorithm that generates a three Degrees of Freedom (DOF) spatial trajectory for an AGV equipped with a 3DOF trajectory tracking controller. The LOS PPG can be used for cooperative, passive (neutral) and adversarial tracking tasks, such as, respectively, formation driving, autonomous lane keeping with speed requirement, and chasing an evading vehicle. The algorithm is verified with computer simulations on a 1/6 scale electric car model, and will be further validated on that model car in the near future.