Real-time path planning for citrus picking robot based on SBL-PRM

A single-query bi-directional probabilistic roadmap planner with lazy collision checking was adopted for citrus picking robot real-time path planning in dynamic and unstructured environments. Simulation of the planner was carried out in two cases of picking exposed and overlapped fruits. The effects of maximum number of milestones S, neighboring threshold ρ, local path checking threshold ε and path smoother steps N on average time and success rate of planning were analyzed. When S, ρ, ε, and N were 3000, 0.6, 0.03, 10 respectively, simulation results indicated that the average planning time was about 1ms for picking exposed fruits and 60 ms for picking overlapped ones, and the success rates of planning were both 100%. The simulation experiment shows that SBL-PRM is effective in the citrus picking robot real-time path planning.