Linas Kondrackis
  • Portfolio

Click on the images to enlarge.

Extended Kalman Filter for position estimation

Computer Vision • C++ • Eigen

The project required to estimate the position of a moving object given noisy Lidar and Radar measurements. It was a part of the Self-Driving Car Engineer Nanodegree in Udacity.

This was a fill-in-the-gaps style task, with a part of the C++ code already given. The tasks involved such tasks as filling in the Jacobian calculation functions, converting from Cartesian to Polar coordinates and allocating the necessary data structures.

Ultimately, I managed to obtain the results that were expected.


Code Repository: GitHub



  • Category: Projects
  • Date: September 2018

A zoom-in to a moment of estimating the path in the second dataset. Red and blue circles are lidar and radar data, and the green triangles are the EKF-estimated path.

The first dataset, without EKF path estimation. Red and blue circles are lidar and radar data.

The second dataset with EKF path estimations. Red and blue circles are lidar and radar data, and the green triangles are the EKF-estimated path.