| Traditional robot planning systems have typically assumed that a perception module could create a complete world model whenever the planning system needed it, or whenever anything in the world changed. This assumption is completely unrealistic in many real-world domains because perception is far too difficult. Robots in these domains cannot use the traditional planner paradigm, but instead must actively reason about what is important to see. In this thesis I describe how reasoning can be integrated with perception, how task knowledge can be used to select perceptual targets, and how this selection dramatically reduces the computational cost of perception.;The domain addressed in this thesis is driving in traffic. I have developed a microscopic traffic simulator called PHAROS that defines the street environment for this research. I have also developed a computational model of driving called Ulysses that defines the driving task. The model describes how various traffic objects in the world determine what actions that a robot must take, and conversely what the robot should look for before making decisions. These tools allowed me to implement robot driving programs that request sensing actions in PHAROS, reason about right-of-way and other traffic laws, and then command acceleration and lane changing actions to control a simulated vehicle.;In the thesis I develop three selective perception techniques and implement them in three robot driving programs of increasing sophistication. The first, Ulysses-1, uses perceptual routines to guide visual search in the scene. The second program, Ulysses-2, decides which objects are important in the current situation. It searches an inference tree to find the most critical sensing actions. Finally, Ulysses-3 uses domain knowledge to reason about how dynamic objects will move or change over time. Objects that do not move enough to affect the robot can be ignored by perception. When run in the PHAROS world, the techniques included in Ulysses-3 reduced the computational cost for perception by 9 to 12 orders of magnitude when compared to an uncontrolled, general perception system. |
