Closed xinjie-liu closed 2 years ago
As I mentioned in the group chat, we can implement a condition in the RRT_star algorithm so that, If the number of iterations is too low so as to find a solution, we can continue iterating until a solution is found or a maximum number of iterations (large, but not too large since we don't want to wait forever) is reached. I will take charge of implementing that next week.
The newly implemented asymptotical optimality radius solves sufficiently well the problem. For a number of iterations bigger than 1500, at least. All implementations I have seen risk termination without finding a solution when too little iterations are given, so I guess it is a common flaw of RRT.
Our RRT planner may should be able to terminate itself once get node close enough to the goal and the corresponding path, we could look into that later as well