Skip to content

benjaminsoellner/CarND_07_UnscentedKalmanFilters

master
Switch branches/tags
Code
This branch is 4 commits ahead, 19 commits behind udacity:master.
Contribute

Latest commit

 

Git stats

Files

Permalink
Failed to load latest commit information.

Self Driving Car Engineer Project 7 - Unscented Kalman Filter

Benjamin Söllner, 31 Aug 2017


Fun Project Header Image


In this project I am utilizing a unscented kalman filter to estimate the state of a moving object of interest with noisy lidar and radar measurements. Passing the project requires obtaining RMSE values lower than the threshold and Normalized-Innovation-Squared (NIS) values within the tolerances outlined by the project rubric.

This project involves the Term 2 Simulator which can be downloaded here.

Screenshot

The process model which describes the motion was tweaked to be comparable to the movement of cyclists (using approx. 0.8m/s² = 2.88km/h per second acceleration change and 22.5 degree/second² angular acceleration).

// Process noise standard deviation longitudinal acceleration in m/s^2
std_a_ = 0.8;
// Process noise standard deviation yaw acceleration in rad/s^2
std_yawdd_ = M_PI / 8.0;

As you can see in the charts below, >80% of the NIS values fall always nicely into the specified range between 0.35 and 7.81:

Chart showing time series with NIS values of lidar & radar measurement

When using both lidar & radar data, 206 out of 249 measurements for lidar (83%) and 226 of 249 measurements for radar (91%) fall within the specified range.

Chart showing time series with NIS values of lidar measurement

When using only lidar data, 209 out of 249 measurements (84%) fall within the specified range.

Chart showing time series with NIS values of radar measurement

When using only radar data, 225 out of 249 measurements (90%) fall within the specified range.

Dependencies

Making

Once the install for uWebSocketIO is complete, the main program can be built and run by doing the following from the project top directory.

  1. mkdir build
  2. cd build
  3. cmake ..
  4. make
  5. ./UnscentedKF

The project can also be built with Microsoft Visual Studio according to a useful article from Fahid Zubair.

Submitted Files

About

Project 7 "Unscented Kalman Filters" of Udacity's "Self Driving Car Engineer" Nanodegree

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • C++ 98.6%
  • C 1.2%
  • Other 0.2%