A Coordinated Global–Local Path Planning Approach for Vineyard Mobile Robots Based on Improved A* and TEB Algorithms
Y F LIU, Jiangchun Chen, Jian Bao, Longpeng Ding, Hongfei Yang, Y F LIU, Yufeng Li, Haiyang Lu, Guangshang Ge · Agriculture · 2026
The semi-structured vineyard environments contain numerous irregular obstacles, posing stringent requirements on the navigational safety and trajectory tracking accuracy of mobile robots. To address this challenge, this study first optimizes the A* algorithm at the global planning layer by incorporating a composite turning-cost evaluation model and a heuristic dynamic weighting strategy, thereby effectively enhancing search efficiency and path smoothness. Building upon this, a local planning method is further developed by integrating an adaptive sampling mechanism with high-order interpolation-based kinematic continuity constraints and a heading-rate-driven velocity smoothing strategy. This enables the robot to maintain a safe clearance from obstacles in dynamic environments, thereby significantly enhancing the smoothness of obstacle avoidance maneuvers. Both simulation and field experiment results demonstrate that the improved global planning algorithm reduces the number of critical turning points and the total turning angle by up to 18.0%. Across three typical path scenarios, the proposed fusion method reduces the robot’s positional deviation by up to 21.8% and the heading angle deviation by up to 29.6%, while concurrently increasing the safe clearance from obstacles by 42.0%. These findings suggest that the proposed framework establishes a viable algorithmic foundation for improving the navigation accuracy, obstacle avoidance stability, and operational safety.