I need to train an ML model to detect when a car in a driving simulator (Carla Simulator) has been overtaken by another vehicle. I realize this can easily be done deterministically but I'd need to demonstrate an ML model predicting this. The data I am getting from the vehicle is in the form of time series data (100 time steps). Each prediction will be made based on 3 parameters describing the position and velocity of the vehicles relative to each other (angle, distance, speed_delta), plus the previous 99 values of these parameters. In total 300 parameter plus the label (overtake y/n).
Currently I have expanded each time step of each parameter into its own column so, I have a dataframe with 300 columns + label. I am thinking of using PCA because of the high dimensionalality of the data but I was wondering if there is anything that may be better suited given the time-series nature of the data.