09. Sample-Based Planning Wrap-Up

Extended Reading

At this point, you have the knowledge to read through a paper on path planning. The following paper, Path Planning for Non-Circular Micro Aerial Vehicles in Constrained Environments, addresses the problem of path planning for a quadrotor.

It is an enjoyable read that culminates the past two sections of path planning, as it references a number of planning methods that you have learned, and introduces a present-day application of path planning. Reading the paper will help you gain an appreciation of this branch of robotics, as well as help you gain confidence in the subject.

Some additional definitions that you may find helpful while reading the paper:

• Anytime algorithm: an anytime algorithm is an algorithm that will return a solution even if it's computation is halted before it finishes searching the entire space. The longer the algorithm plans, the more optimal the solution will be.

• RRT*: RRT* is a variant of RRT that tries to smooth the tree branches at every step. It does so by looking to see whether a child node can be swapped with it's parent (or it's parent's parent, etc) to produce a more direct path. The result is a less zig-zaggy and more optimal path.