[MT] Human-based Constraints for Driverless Go-karts #1030
Description
Motivation
Apart from racing purposes, the go-kart is usually considered a ludic device that can be driven by all kind of possible users (children, young people, adults, old people), that are generally characterized by different driving styles and different perception of risks. In this thesis, we want to implement human-based constraints about path/accelerations/collision speed in the pre-existing MPCC strategy, aiming both at ensuring safe trajectories, both at improving driving comfort and fun. As an example, strong accelerations should be penalized when a child is driving; on the contrary, an adult might be more excited by high accelerations and strong braking.
Tasks
- Definition and implementation of human-constraints in the preexisting MPCC in terms of path/speed/accelerations.
- Definition of suitable cost functions with different aims (driving comfort, safety).
- Introducing the possibility of low-speed collisions with obstacles on the track.
- Providing the user with the authority of controlling the pedals, in a cooperative driving framework.
Expected outcome
Appropriate documentation of all the experiments showing the ability to turn good theoretical models into working code. Modular and reusable implementation of the tasks.