home
past talks
resources
conferences
|
|
|
Opportunities for Machine Learning in Path Planning
Brendan Burns
UMass
Abstract
Path Planning in high dimensional spaces is a known difficult problem. At the same time, there are a number of interesting applications (humanoid robots, protein folding) which require us to be able to path plan in high dimensional spaces. An approach to solving this problem is a randomized technique known as probabilistic road-map construction.
I'll briefly present background on the probabilistic road-map algorithm and then explore ways in which we have begun to use traditional machine learning techniques to attack the problem.
Back to ML Lunch home
|