Before the steering angle of a parking car can be determined, a simplified model has to be created. The Ackerman steering model is a well known example which contains of 4 wheels and the ability to change the direction. This can simulate a car movement in all directions.
The open problem is how to steer the angle to fulfill a condition. For example to park the car into a lot. A weak technique in doing so is a fuzzy controller. The idea is to connect natural language action words with a steering angle and then the policy is created by the inference engine. The problem with fuzzy control is, that the reality is working with a different principle. That means, the calculated trajectory isn't accurate enough. The better idea over a fuzzy controller is a neural network which can adapt to new situations.
- Vigneshwaran, R., et al. "Optimization of Four Wheel Steering." Int. J. of Multidisciplinary and Current research 6 (2018).
- Firdaus, AZ Ahmad, et al. "Design and Simulation of Fuzzy Logic Controlled Car Parking Assist System." Journal of Telecommunication, Electronic and Computer Engineering (JTEC) 10.1-15 (2018): 81-87.