teadetime / crazyflie-path-planning Goto Github PK

Add checks to shape of parameters and adjust accordingly.

Goals for a design review on Friday 4/21:

• Solve for position waypoints for navigating a single agent around a single square box (L1 ball) in jupyter notebook sandbox
• Solve for (position? velocity?) waypoints for navigating two agents around a single square box
• Add support for multiple obstacles in sandbox
• Integrate with broader integrated codebase (possibly for next sprint)

Spec for what we plan to have done by DR1

• Robust tooling for nice future development
• Some form of visualization of occupancy grid
• Visualization of multi-agent A* based algorithm
• List of other algorithms to use (papers read on these/feasibility check)

See #1

• Algorithms should be broken into tested sub-functions when applicable
• Every function should be typed
• Every file,class, and function should have a doc string (description and args)
• Have a standard test file for planned paths

We need to support the additon of cost for a given vertex? should this be a duplicated occupancy grid? or simply multiply the dimensions and change datatypes so everything is a float or NAN

Nathan -> Proposes multiple arrays with helper functions to get and set cost for lists of , this allows the original map to stay boolean.
Adding dimensions could be helpful though, and our queries might get a little hairy. I am pro multiple arrays and indexing.

Change function signature and create a state class.