SFND_Unscented_Kalman_Filter

Sensor Fusion UKF Highway Project Result

Meaning of colors and shape

• Green Car : Ego Vehicle
• Blue Car : Ground Truth of Target Vehicle
• Red Sphere : Lidar Sensing Data (px, py)
• Purple Line/Arrow : Radar Sensing Data (range, angle, doppler)
• Green Sphere/Arrow : UKF Result (px, py, v, yaw, yaw_rate)

---

Dependencies

• cmake >= 3.5
• make >= 4.1
• gcc/g++ >= 5.4
• PCL 1.2
• Or Use my docker image : docker run -p 6080:80 -v /dev/shm:/dev/shm kimjw7981/sfnd

Basic Build Instructions

1. Clone this repo.
2. Make a build directory: mkdir build && cd build
3. Compile: cmake .. && make
4. Run it: ./ukf_highway

Step By Step

1. Turn Off all sensors
   
   RMSE Fail: X, Y, Vx, Vy
   Reason: No sensing data and keep initial values

2. Predict and use only LiDAR data (KF) for updating with measurement
   
   RMSE Fail: Vx
   Reason: LiDAR can detect position directly and isn't used for calculating velocity

3. Predict and use only Radar data (UKF) for updating with measurement
   
   RMSE Fail: Y, Vy
   Reason: Radar can detect position and velocity with low resolution.

4. Predict and use Sensor Fusion data with LiDAR (KF) and Radar (UKF) for updating with measurement
   
   RMSE Fail: None
   Reason: Sensor Fusion data makes better result (Reduce each other's errors)