Hello,
Although, I´ve done some research on the topic, I am still relatively new to Trajectory planning. Hopefully someone can help me with the problem, that I am currently having.
I would like to create a simplified simulink model for a Trajectory planner, which would allow an ego-vehicle to follow a simple refence path (The reference path is theoretical, which means, that it´s coordinates are initially not known). The vehicle could be presented as a bicycle model with a constant speed (low speed).
At this moment I consider to use MPC(Model predictive control) similar to: https://www.youtube.com/watch?v=4IO1zxsY4wU , but without the visualisation. As in the visual representation, the vehicle is not per se placed on the path and is only able to recognize the path at a certain distance. Thus the trajectory planning should only be able to compute a feasible trajectory within a certain distance. Following constraints should be considered:
- Obstacle recognition and avoidance (recognition at same distance, as path recognition and trajectory replanning)
- Bicycle Model constraints (maximum steering angle)
- (secondary) Avoidance of jerky movements
Following inputs/outputs should be implemented:
- inputs: the vehicle state (X,Y, Yaw, Yaw rate) , reference Path state (X, Y, curvature)
- outputs: yaw, yaw rate
I am very thankful for any helpful advice on this topic!
Best Regards!
