Project Ideas
- You are encouraged to develop your own topic, but here are a few examples:
-
Develop a motion planner for an aerospace vehicle required to inspect a constrained 3D environment. You might consider a spacecraft with thrusters inspecting a space satellite for cracks or defects, a helicopter searching for disaster victims in an urban setting, or a submarine homing on the source of a plume near the ocean floor.
- Develop a planner for one or more aerospace vehicles to find an evading target. (Also develop a planner for the target!) You might consider a situation where the pursuer does not want to be detected either (the "private detective" problem in city blocks).
- Develop a planner for an aerospace vehicle that is better integrated with a low-level controller. For example, here is a typical way this integration is performed: first, we plan a motion; then, we give it to the controller for execution; finally, if anything goes wrong, we plan a new motion from the current position. But what is the definition of "wrong"? How much does the vehicle's actual trajectory have to deviate before we re-plan? You might design a motion planner that computes the amount of deviation allowable before re-planning.
- Develop a planner for a car (or planetary rover) driving on ice.
- Develop a planner for an underwater vehicle tracking a jellyfish (or other marine life) that automatically performs tether maintenance while keeping the target in view. You might even consider multiple underwater vehicles.
- Develop a planner for an unmanned aerial vehicle flying through a forest.
- Develop a planner for a pure glider (that must use thermals, for example, to remain aloft).
- Identify an unstudied aerospace vehicle that is underactuated or subject to non-holonomic constraints, prove that it is (or is not) controllable, and design a motion planner for it.
- Implement a sampling-based verification scheme for air traffic control or a similar safety-critical system of aerospace vehicles.
- For a rover with a wheels-on-legs design, develop an algorithm that quickly decides which mode of locomotion is best to use, given local sensor information.
- Develop a systematic approach to designing a good set of maneuvers for an aerospace vehicle (or a group of vehicles). Demonstrate this approach by developing a motion planner for this vehicle.