Key points are not available for this paper at this time.
Reliable autonomous navigation under canopy occlusion, dust, uneven terrain, and moving machinery remains the central technical barrier to scalable farm robotics. Achieving this reliability requires integrated systems that combine robust perception, precise localization, dependable obstacle detection, efficient path planning, and accurate motion control. Based on 68 peer-reviewed articles from 2015 to 2025, this review categorizes agricultural environments into structured, semi-structured, and unstructured environments, highlighting how corridor regularity, obstacle dynamics, terrain variability, and localization reliability affect navigation performance. This review discusses various path planning techniques, types of obstacles, and strategies for obstacle avoidance and localization. It emphasizes the integration of sensor fusion from light detection and ranging (LiDAR), cameras, radar, and other sources to enhance real-time decision-making and navigation precision. In addition, the paper reviews mapping and localization techniques, including simultaneous localization and mapping (SLAM) and multi-sensor fusion, while addressing challenges in mapping accuracy, localization, power management, and safety to support advanced agricultural production. A future approach could be the integration of temporal data into the path planning algorithm for robust and adaptive navigation. Overall, this review provides a comprehensive assessment of path planning techniques in agricultural settings and analyzes the capabilities and limitations of different sensor systems for obstacle detection and localization.
Singh et al. (Wed,) studied this question.