ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange |
1 | initial version |
I answer myself: increasing weight_optimaltime from 1 (default) to 10 makes this behavior much rarer. And if still happens (and the robot gets trapped oscillating), reducing force_reinit_new_goal_dist to 0 will make TEB to discard intermediately the faulted plan and start anew (though to my experience this can only be done as a recovery mechanism, as it makes TEB very unstable)