The scenario is composed by a straight highway with three lanes and three actor car, besides the ego car. The ego car is green while the actor cars are blue. The viewer scene is centered around the ego car and the coordinate system is relative to it.
The actor cars accelerate and change lanes as the ego drives down the road. Each of actor cars has it’s own UKF object which is independently updated at each time step.
The red spheres above cars represent the (x,y) lidar detection and the purple lines show the radar measurements with the velocity magnitude along the detected angle. Altitude not taken into account for tracking, so this is only tracking on a horizontal plane parallel to the road.
In this project an Unscented Kalman Filter is implemented to estimate the state of the actor cars on a highway using noisy lidar and radar measurements. As part of the goal, the RMSE values must be
The simulation collects the position and velocity values that the algorithm outputs and they are compared to the ground truth data. The RMSE values for px, py, vx, and vy RMSE should not exceed [0.30, 0.16, 0.95, 0.70] respectively.
The Unscented Kalman Filter is implemented using mostly functionality developed during the quizzes performed during lessons such as Augmentation Assignment, Predict Sigma Points and Predict Mean and Covariance. There was a bit of trial and error setting the parameters for longitudinal acceleration noise and yaw acceleration noise but just a bit of reasoning around expected values is sufficient to get the desired RMSE overall.