Closed rayvburn closed 1 year ago
Local minima avoidance is executed via min_force
parameter, which, when set to >1 value, is able to prevent from being stuck.
On the other hand, local minima problem also affects movement when goal is placed near some obstacle - robot oscillates when close to the goal. The solution is to decrease interaction factor when approaching to the goal.
ROS base_local_planner
has recovery behaviours to recover robot being stuck in 1 place for a longer time
Currently local minima is avoided with a use of Simple Sampling Based Trajectory Generator that aims to find a feasible control even when the robot cannot generate reasonable model-based trajectories. Recovering trajectories still must be properly scored by critics before they will be used to command the mobile base.