A unified adaptive pure pursuit speed controller with EKF Sensor Fusion for real-world ackermann mobile robots

Abstract

Accurate trajectory tracking is critical for autonomous mobile robots navigating in outdoor environments. This research presents an enhanced version of the Pure Pursuit algorithm, referred to as Pure Pursuit with Dynamic Steering Control (PP-DSC), which modulates the robot's velocity according to the magnitude of the steering angle while maintaining a fixed lookahead distance. The algorithm was implemented on a four-wheeled robot with Ackermann steering, and localization was achieved by fusing data from the Global Navigation Satellite System Real Time Kinematic (GNSS-RTK), Inertial Measurement Unit (IMU), and encoder data using an Extended Kalman Filter (EKF). Field experiments were carried out on three representative paths: straight line, S-curve, and loop at operational speeds ranging from 1.0 to 3.0 m/s. The results demonstrated that PP-DSC consistently reduced lateral deviation compared to fixed-speed Pure Pursuit. Root Mean Square Error (RMSE) was reduced to 0.9 cm on the straight-line path, and on the S-curve path, RMSE was reduced to 2.1 cm. The RMSE decreased to 1.9 cm for the loop path with PP-DSC, while the 1 m loo kahead configuration exhibited a higher RMSE of 2.7 cm. These findings confirm that steering-based velocity modulation effectively improves path-tracking precision under real-world outdoor conditions.