Experimental investigation of automatic operation trajectory planning and tracking of unmanned excavators

Abstract

Excavator automation is essential for intelligent construction machinery but faces challenges like unstable joint movements and impact forces. To address this, this study presents a trajectory planning and control method combining robotic trajectory planning with adaptive pure pursuit control. First, the kinematic model of excavator working mechanisms is established using Denavit–Hartenberg (D-H) parameters, with excavation trajectories generated through cubic B-spline interpolation of key operational points extracted from actual digging patterns. A piecewise polynomial-based trajectory planning method ensures smooth bucket motion by optimizing joint transitions. Subsequently, an adaptive pure pursuit controller with dynamic look-ahead distance adjustment is developed for path tracking. Finally, experimental validation on an 8 ton electric excavator demonstrates effective performance: trajectory tracking achieves mean absolute errors of 19.7 mm (slope leveling) and 16.6 mm (excavation), with corresponding root mean square errors of 26.3 and 23.4 mm. These results confirm enhanced joint movement smoothness while maintaining sub-30 mm operational precision, fulfilling autonomous excavation requirements through coordinated path planning and adaptive control integration.