Changes to nav2 planner and controller if required. #375
Description
Summary
We're currently using a simple controller, but there are a lot more options. MPPI controller is currently popular, is probably a good fit. Requirement is that the controller and planner a so tested, standard and stable that after we finished the navigation milestone we don't need new planners / controllers or do rework for at least 6 months.
When going from waypoint to waypoint with a robot platform in simulation, the path can vary from iteration to iteration. This causes inconsistent behaviour, which could be dangerous in public spaces. Perhaps different parameters or a different planner is needed.
When planning, the robot can switch between two paths indefinitely:
DoD
- Check which controllers are available, describe pros and cons and choose one.
- Test and try for different controllers shortly to validate assumptions
- Install new controller
- Check which planners are available, describe pros and cons and choose one.
- Test and try for different planners shortly to validate assumptions
- Install new planners
- Test in simulation
- When possible, test on robot
- Update documentation