Self-Driving Car Engineer Nanodegree Program
In this project, I built an Unscented Kalman Filter in c++ to fuse sensor data from 2 sources:
- LIDAR Sensor
- RADAR Sensor
A constant turn rate and velocity magnitude (CTRV) model is used to predict the states.
There are sample data files in the /data folder.
Run the UKF in a terminal entering:
./UnscentedKF /input/data/file /output/data/file
The following plots were made using the jupyter notebook ukf-visualisation-extended.ipynb (Taken from here and adjusted to my needs).
The process noise parameters std_a_ and std_yawdd_ were tuned using Normalized Innovation Squared (NIS) Values
The results are reasonable. See the RMSE Values in the table below.
| Value | RMSE |
|---|---|
| x | 0.07 m |
| y | 0.08 m |
| vx | 0.34 m/s |
| vy | 0.23 m/s |
The plots comparing ground truth and UKF estimations support the low RMSE values.
X vs Y Position:
Absolute velocity vs time:
Yaw angle vs time:
Yaw rate vs time:
