top of page

Robotics

Project | Camera feature tracking (C++)

Implemented a pipeline for tracking a vehicle using 2D camera images with multiple detector/descriptor combinations using OPENCV C++ libraries. A number of traditional and binary detectors/descriptors were implemented in the pipeline for evaluation.

Project | LiDAR Object Detection (C++)

Implemented a pipeline for object detection in LiDAR point dataset. The pipeline reduces LiDAR points using voxel grid reduction for faster detection, followed an implementation of RANSAC to separate ground LiDAR points from object data points. The separated object points are clustered together using a kd-tree structure. The cluster point clouds are detected objects.

Project | Vehicle Tracking using Sensor Fusion and Unscented Kalman Filter (C++)
The pipeline implements UKF to track a vehicle moving on a non-linear path forming an eight shape in a simulator. The pipeline has been implemented in C++. Simulated LiDAR and Radar data is used by the filter to estimate and track the car in the simulation.
Project | Vehicle Tracking using Sensor Fusion and Unscented Kalman Filter (C++)
The pipeline implements UKF to track a vehicle moving on a non-linear path forming an eight shape in a simulator. The pipeline has been implemented in C++. Simulated LiDAR and Radar data is used by the filter to estimate and track the car in the simulation.
Project | Prey-Predator Hunting using Extended Kalman Filter (Python)

Implemented a pipeline of Extended Kalman Filter and planning logic for a simulated predator to hunt a prey moving at a speed equal to or higher than the predator. The green dot is the prey and the blue is the predator.

giphy.gif
Project | Object tracking in 3D space (C++)

Implemented a pipeline for tracking and calculating Time-To-Collision with preceding vehicle using LiDAR and Camera data in 3D space. The LiDAR data points have been projected on to the vehicle camera image frame and a bounding box created to encasulate preceding vehicle.

Project | Vehicle Tracking using Sensor Fusion and Extended Kalman Filter (EKF) (C++)
The pipeline implements EKF to track a vehicle moving on a non-linear path forming an eight shape in a simulator. The pipeline has been implemented in C++. Simulated LiDAR and Radar data is used by the filter to estimate and track the car in the simulation.
Project | PID Controller for Self Driving Car (C++)

A PID controller is implemented in C++ for a self-driving car with the controller goal being to minimize the Cross Track Error (CTE) for the car. The controller was tested in a UNITY engine based simulation environment provided by UDACITY. 

Project | RADAR Target Generation and Detection (MATLAB)

A project simulation RADAR target generation and detection in MATLAB. A simulated FMCW signal from a target is detected and processed in a pipeline to determine the range and velocity of the vehicle using 2D FFT and CFAR.

Project | MPC Controller for Self Driving Car (C++)

A self driving car will have drive itself based on control inputs to minimize the Cross Track Error (CTE) for the car. The implementation is done in python. This project completed as part of an online Udacity course and the Self driving car simulator provided by Udacity is used to test the implementation

bottom of page