/EKF

Extended Kalman Filters in C++

Primary LanguageJupyter NotebookMIT LicenseMIT

EKF -Extended Kalman Filters in C++

Udacity CarND Term 2, Project 1 - Extended Kalman Filters

Project Basics

The main goal of the project is to apply Extended Kalman Filter to fuse data from LIDAR and Radar sensors of a self driving car using C++. The code will make a prediction based on the sensor measurement and then update the expected position.

See the starter code from Udacity Project Starter Code

Basic Build Instructions

  1. Clone this repo.
  2. Make a build directory: mkdir build && cd build
  3. Compile: cmake .. && make On windows, you may need to run: cmake .. -G "Unix Makefiles" && make
  4. Run it: ./ExtendedKF

Contents of Repository

  • src dir
    1. main.cpp - communicates with the Term 2 Simulator receiving data measurements, calls a function to run the Kalman filter, calls a function to calculate RMSE
    2. FusionEKF.cpp - initializes the filter, calls the predict function, calls the update function
    3. kalman_filter.cpp- defines the predict function, the update function for lidar and the update function for radar
    4. tools.cpp- function to calculate RMSE and the Jacobian matrix
  • data - a directory with one input file, provided by Udacity
  • results - a directory with output and log files
  • Docs - a directory with files formats description

Results

Using two sets of simulated runs, Extended Kalman Filter produces the below results. The x-position is shown as 'px', y-position as 'py', velocity in the x-direction is 'vx', while velocity in the y-direction is 'vy'.

The graphs are created below using ekf-visualization.ipynb provided by Mercedes Sensor Fusion Utilities This is done by logging the data in an output.txt file when running the prject with simulator. Format of the logged data is prsenet in following sequence: ['px_est','py_est','vx_est','vy_est','px_meas','py_meas','px_gt','py_gt','vx_gt','vy_gt']. This is then observed using ekf-visualization.ipynb.

Threshold RMSE <= [.11, .11, 0.52, 0.52].

1. Laser-Radar Combined

Set following compiler switches to:

ONLY_LASER 0

ONLY_RADAR 0

1. Dataset 1

Laser-Radar-Output-Dataset1

Accuracy - RMSE [0.0974, 0.0855, 0.4517, 0.4404] Accuracy - RMSE

ekf_output1.txt

2. Dataset 2

Laser-Radar-Output-Dataset2

Accuracy - RMSE [0.0726, 0.0967, 0.4469, 0.4944]

Accuracy - RMSE

ekf_output2.txt

2. Laser Only

Set following compiler switches to:

ONLY_LASER 1

ONLY_RADAR 0

Laser-Output-Dataset1

Accuracy - RMSE [0.1222, 0.0984, 0.5825, 0.4567]

ekf_laser_output1.txt

2. Dataset 2

Laser-Output-Dataset2

Accuracy - RMSE [0.0961458, 0.100295, 0.541759, 0.464002]

ekf_laser_output2.txt

3. Radar Only

Set following compiler switches to:

ONLY_LASER 0

ONLY_RADAR 1

1. Dataset 1

Radar-Output-Dataset1

Accuracy - RMSE [10.9958, 7.7916, 10.109, 7.8036]

ekf_radar_output1.txt

2. Dataset 2

Radar-Output-Dataset2

Accuracy - RMSE [0.224622, 0.295356 0.566935, 0.724152]

ekf_radar_output2.txt

Observation

From the above results (RMSE values), Laser data seems to provide more accurate readings than the radar. Combining both of them strengthens the EKF algorithm to predict the position more precisely.