Comparative Study on Simulated Outdoor Navigation for Agricultural Robots.

This research presents a comprehensive comparative analysis of SLAM algorithms and Deep Neural Network (DNN)-based Behavior Cloning (BC) navigation in outdoor agricultural environments. The study categorizes SLAM algorithms into laser-based and vision-based approaches, addressing the specific challenges posed by uneven terrain and the similarity between aisles in an orchard farm. The DNN-based BC navigation technique proves efficient, exhibiting reduced human intervention and providing a viable alternative for agricultural navigation.