- Task
- The most important part of a motion planner is the problem it solves.
- There are 4 tasks for the purposes of this book,
- navigation
- coverage
- passing a sensor or tool over all points in space, such as painting
- localization
- using a map to interpret sensor data
- mapping
- exploring and sensing an unknown environment
- Properties of the Robot
- Degrees of freedom
- Configuration space
- Consider the shape - is the robot free to move instantaneously in any direction in its configuration space? If so, it is omnidirectional. if it is subject to velocity constriants (like a car that can’t translate sideways), it is nonholonomic.
- Properties of the Algorithm
- does a planner find optimal solutions for properties like length, execution time, or energy consumption?
- computational complexity of the planner
- complete - meaning they wil always find a solution to the planning problem when one exists or indicate failure in finite time
- optimallity, completeness, and computational complexity naturally trade off with one another.
- planners are offline if they construct a plan in advance, online if it incrementally constructs the plan while in execution.