Open jeking04 opened 9 years ago
You can try switching back to CBiRRT or OMPL for our default randomized planner. I switched to the OpenRAVE BiRRT based on the input of @siddhss5 and @psigen. They found that the OpenRAVE BiRRT is much faster than OMPL's RRT-Connect.
I've found that this pair of errors generally means that the planner thinks the initial configuration is violating a limit or in collision.
On Mon, Apr 6, 2015 at 9:29 PM, Michael Koval notifications@github.com wrote:
You can try switching back to CBiRRT or OMPL for our default randomized planner. I switched to the OpenRAVE BiRRT based on the input of @siddhss5 https://github.com/siddhss5 and @psigen https://github.com/psigen. They found that the OpenRAVE BiRRT is much faster than OMPL's RRT-Connect.
— Reply to this email directly or view it on GitHub https://github.com/personalrobotics/prpy/issues/101#issuecomment-90306404 .
It could be the usual "padded joint limits" problem, like @jeking04 suggested, then. Maybe it's time to actually solve this properly...
I see this error often when planning via Birrt:
[rrt.h:80 _InitPlan] no initial configurations [rrt.h:354 PlanPath] BirrtPlanner::PlanPath - Error, planner not initialized
It is not consistent, i.e. sometimes we are able to plan just fine. When I manually move the arm to reset we seem to be fairly wrapped up, so it could be because the start configuration is near joint limit.