![]() ![]() If a faster vehicle needs to overtake a slower one, the slower one is paused so that vehicles don’t have to handle dynamic passing.Must follow speed limits for specific areas of the course to protect infrastructure and ecology.Course kept secret until 2h before the race. ![]() Must traverse a desert course up to 175 miles long in under 10 h.In 2005, 195 teams registered, 23 teams raced, 5 teams finished.No one made it through more than 5% of the course in 2004 race.Goal: to build a vehicle capable of traversing unrehearsed off-road terrain.Defense Advanced Research Projects Agency (DARPA).Actions could be nondeterministic depending on data (or lack thereof).Often require simultaneous planning and execution (online).Planning can frequently be done offline.Unfortunately, these are very high expectations which don’t always correlate very well with machine learning techniquesĭifferences Between Other Types of Machine Learning and Robotics Other ML Applications Robotics.Learn through their experience and human guidance.Work in situations where information about the environment is imperfect or incomplete.Be able to react promptly and correctly to changes in environment or internal state.Monitoring actions to ensure correctness.Learning to interact with the world and with objects.Example: robot navigating unknown environment in search of a goal.Policy: generating correct actions to reach the goal.Clustering: find groupings in the input.Density Estimation: finding patterns in the input space.Want to find regularities in the input.Input: data from sensors (like GPS or video).Training Data: actions of human drivers in various situations.Both Classification and Regression are “Supervised Learning” strategies where the goal is to find a mapping from input to output.Outlier Detection – finding exceptions to the rules.Knowledge Extraction – learning a rule from data. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |