Design of Vehicle Dynamics feature set

[adamdbrw]

The goal is to create Components and classes to support simulation of movement of diverse types of robot mobile bases. The focus in this task should be narrowed down to the type required by the MS2 demo (a four-wheeled, Ackermann drive type), with only rudimentary analysis of other types to understand what generalizations might be useful for the design. This task should include the following items:

• [x] Domain research, including available open-source solution (PhysX vehicle Gem, others) and solutions in other simulators.
• [x] Design of components and classes. How would a developer use the feature?

Outcomes:

• [x] Document capturing useful knowledge of how this feature is designed in existing solutions, comparison of major advantages and weaknesses.
• [x] First design document
• [x] Accepted design document, after a review.
• [x] Finally, a set of issues capturing implementation of the proposed design (some issues can be outside of MS2).

[adamdbrw]

VehiclePhysics.pdf First draft for reading. This will be converted to a PR with docs

[adamdbrw]

First iteration has been submitted. Sub-tasks will include:

• Tasks for implementation of proposed components
• Tests on physics scene
• Implementation of a working example for a four-wheeled robot
• Second iteration of design

[pawelbudziszewski]

It seems fine to me, I have just one question regarding the diagram: why the WheelControllerComponent is connected to the VehicleModel via WheelsConfiguration, not directly? I'm not saying this is wrong, it just seems more natural for me if these wheel-related components are swapped places.

[adamdbrw]

WheelsConfiguration is a part of VehicleModel which holds (references) to WheelControllerComponents.

[pawelbudziszewski]

Ok for me