- Radar/Lidar tracking- Sensor fusion- Performance evaluation

I was part of the sensor fusion team under the supervision of Dr. Ting Yuan in the autonomous driving division of MBRDNA. I have worked on the following projects from algorithm development to code implementation on the testing car.1. Unbiased Interacting Multiple Model (IMM) estimation for extended (kinematics and geometry) object tracking: Algorithms: Kalman filter, IMM (models: constant velocity, acceleration, turn, coordinated turn, stop model), Data association (Single Nearest Neighbor, Global Nearest Neighbor, Probabilistic Data Association (PDA), IPDA, Geometric intersection), Gating (distance, Mahalanobis, speed), Sliding window approaches, Track existence/ management, initialization techniques, coordinate transformation (ego, UTM, local-stationary)2. Unbiased Interacting Multiple Model - Sequential Multiple Hypothesis Test (IMM-SMHT) for object matching to improve the association/ estimation accuracy in inter-vehicle communication system (manuscript ready for journal submission)Sub-projects:3. Principal component analysis based extended object tracking 4. Cramer-Rao Lower Bound (CRLB) module for multi-target multi-sensor scenario5. Time-slicing in multiple sensor scenario, static grid map 6. Localization based on GPS data7. Feature estimation of an extended object from Doppler radar (DBSCAN, RANSAC, velocity-profile information) Other: Radar/ Lidar/ Stixel data processingTechnologies: C++, ADTF (Automotive Data and Time-Triggered Framework), Ubuntu, Matlab, Sconstruct, CMake

Successfully developed and tested a C++ module for vessels using the historical AIS information in maritime environment. Historical AIS information collected for few months were used to create the historical shipping lane information. A grid based landmass avoided maritime map was created using GSHHG (A Global Self-consistent, Hierarchical, High-resolution Geography Database) database.The module provides the following functionalities:1. Landmass avoided routing (grid based approach - not image based)2. Path prediction - Algorithms used: Dijkstra, A* 3. Estimated Time of Arrival (ETA)The algorithms I have developed during my Master's thesis were included in this module as well. Technologies: C++, PostGIS, PostgreSQL, pgRouting, CentOS, GSHHG, AIS, Jenkins, ConfluenceThe module will be integrated as a service in the product level soon.Supervisor: Dr. Bal Randhawa

A research assistant at the ETF research laboratory of the Electrical and Computer Engineering department of McMaster University. Currently, I am working in collaboration with other members of the Estimation Tracking and Fusion (ETF) lab to develop extended object tracking based on automotive radar data in autonomous vehicle systems. The project involves research related to multiple motion model tracking, short range radar detection modeling, extended object representation, data association techniques for extended objects, clustering algorithms for radar data processing (DBSCAN, RANSAC), velocity profile analysis of radar detection etc.

Linear Systems (ECE 707)Signals and Systems (ECE 3TP4)Introduction to Linear Control Systems (ECE 3CL4)Control Systems Design (4CL4)Electronic Devices and Circuits I (2EI5)*Supervised and evaluated lab projects, assignments and exams*Conducted tutorials for groups of approximately 100 students*Holding weekly office hours to help students through their homework problems

Worked in a team as a developer for projects in .NET framework and Windows Presentation Foundation (WPF)

Worked in following divisions*Radio Network Planning(RNP) & Radio Network Quality (RNQ)*Transmission planning & Operations*Switch planning & Operations*Value added services(VAS) planning & Operations*Network Operations & International Network Planning