What is motion planning?
Motion planning figures out how a robot should move to accomplish a goal without hitting anything or violating its own physical limits. Given a start configuration, a goal, and a model of the environment and the robot, it searches for a trajectory through the space of possible configurations that is both collision-free and feasible to execute.
It is the bridge between deciding what to do and actually doing it in the physical world.
Key takeaways
- Motion planning finds a feasible, collision-free path to a goal.
- It reasons over the robot's configuration space and constraints.
- It connects high-level goals to executable movement.
How it works
The planner represents the robot's possible states as a configuration space and searches it for a valid path, using methods such as sampling-based planners that randomly probe the space or graph-search methods over a discretized map. The resulting path must respect obstacles, joint limits, and dynamics, and it is then handed to a controller that executes it. Perception, including maps from SLAM, supplies the obstacle information the planner needs.
Why it matters
Motion planning is essential wherever a robot moves through or manipulates the world, from a mobile robot crossing a room to an arm reaching around clutter. It ties perception and control together, and increasingly it is paired with or replaced by learned policies for complex tasks.
Frequently asked questions
What is a configuration space?
It is the space of all possible poses or joint settings of a robot, where planning searches for a valid path from start to goal.
How does motion planning relate to control?
Planning computes the path to follow, while control produces the low-level commands that make the robot actually track that path.
Related terms