Deep Learning-Based Autonomous Excavation: A Bucket-Trajectory Planning Algorithm

The increased risk to the safety of excavator personnel and difficulty in training them, combined with a manpower shortage, have led an demand for machine automation. This study applies long short-term memory algorithm automating bucket-tip trajectory planning AI system. Unlike other autonomous excavation techniques, proposed approach this performs bucket-trajectory without prior knowledge nonlinear bucket–soil interaction dynamics during excavation, which requires precise adjustment parameters heuristic analysis correlation between them. Based on data acquisition from experts, method uses three-dimensional point cloud terrain bucket motion process application AI. Especially, we transform cloud, comprises massive number points increases computation complexity, into much smaller values, are sufficient representing shape target terrain. To ensure against collision underground obstacles, avoidance is applied prevent crashes excavations along given path, based continuous monitoring pressure excavator’s hydraulic cylinder. Comparative experiments reveal that system generates traceable controller, equipped excavator, yields desired volume lead time collision, regardless topographic change caused through successive excavations.