Foundation construction of an operation support system for outdoor work machines using deep reinforcement learning

It is important for the safety of outdoor machines such as crawler cranes to avoid hazardous event such as tipping, structural damage, and excessive motion under the influence of wind and the ground supporting the machine, which is the natural environment. We are focusing on lifted load sway, irregular excessive motion occurring during materials handling operations in crawler cranes due to factors like wind, and are developing the foundation for an operation support system using deep reinforcement learning. This work incorporates a simulator integrated with the 3D game engine Unreal Engine 5. In this study, we describe the load sway suppression performance achieved by the deep reinforcement learning agent created in MATLAB/Simulink, as well as its interoperability with human operation.