Summary
As my first project during my master’s in Control and Automation, my ground and I got to work with a skid-steered vehicle called the RobuROC. The goal of the project was to explore the usage of wheel slip estimation to increase the performance of an autonomous self-driving car. In order to do this, we had to develop the entire control system for a self-driving car, this included:
- Rewiring of the entire car, as the vehicle was in a bad state
- Modeling of the car using Newtonian equations (and state space)
- Design of an optimal controller
- Path following
- Kalman filtering for better navigation (and slip estimate)
- Lots of software
In the end, it turned out pretty good, and we found that the usage of the estimated slips could be used to increase the system performance. A video showcasing our system in action can be seen here: