Autonomous Path Finding and Obstacle Avoidance Method for Unmanned Construction Machinery

The working environment of construction machinery is harsh, and some operations are highly repetitive. realization intelligent helps to improve economic efficiency promote industrial development. Construction different from ordinary passenger vehicles. Aiming at the fact that existing environmental perception data set cannot be directly applied machinery, this paper establishes corresponding in combination with specific conditions carries out training based on PointPillars network realize function applicable machinery. Most runs unstructured roads, vehicle path planning algorithm not Based this, uses a hybrid A* achieve meets kinematics realizes real-time obstacle detection avoidance. At same time, combines provide method autonomous finding avoidance for improved pure pursuit algorithm, high-precision motion control established trajectory tracking realized, which lays certain foundation follow-up research development related technologies