kalman

Unscented Kalman filter (UKF) C++ implementation based on Udacity Nanodegree Programm "Sensor Fusion Engineer"

Project content

Kalman filter are used in this project to estimate positioning and velocity and yaw angle/rate using lidar or/and radar measurement data.

• An unscented Kalman filter is implemented in ukf.cpp and ukf.h
• tools.cpp/.h controls the ukf filter and its visualization
• highway.cpp/.h define an example scenario for testing

Installation

git clone https://github.com/schottb85/kalman.git

cd kalman

mkdir build

cd build

cmake ..

make

Variants

Visualization options

There are different flags to be selected for different visualization in highway.h

• bool visualize_lidar = true;
• bool visualize_radar = true;
• bool visualize_pcd = false;

Lidar and/or Radar UKF measurements

• in ukf.cpp two parameters bool use_laser_ and bool use_lidar_ can be used activate/deactivate the use of lidar and radar measurements, respectively

Prediction of path

the path prediction for the different cars can be configured using e.g.

• double projectedTime = 2;
• int projectedSteps = 6; defining the prediction/projection time in seconds and the number of discrete ukf projection steps