My research interest lies in estimation, target tracking, pedestrian motion modeling, dependent targets motion modeling, prediction and smoothing. and I have been working on these areas since I joined ETF lab in September 2010. During my master's degree I worked on research related to path-constrained targets. My master's thesis title is "Prediction, tracking, and retrodiction for path-constrained targets".


  • Tracking interdependent target motion using the PHD filter
Signal Processing, Sensor/Information Fusion, and Target Recognition XXV, April 2016.

  • The Social Force PHD Filter for Tracking Ground Targets
Submitted for IEEE Transactions on Aerospace and Electronic Systems, 2014.
  • Social force PHD filter for ground targets
Abstract presented in Technological Advances in Science, Medicine and Engineering (TASME), Toronto, July 2014.
  • SMC Social forces PHD filter for dependent ground targets
Poster presentation, McMaster University, April 2014.

  • Prediction and retrodiction for path-constrained targets with low revisit rate senors
IEEE Transactions on Aerospace and Electronic Systems Journal, vol.50, no.4, pp.2746, 2761, October 2014.

  • Maritime vessel tracking
Abstract presented in Technological Advances in Science, Medicine and Engineering (TASME), Toronto, July 2013.

  • Prediction, tracking, and retrodiction for path-constrained targets
Presented at Canadian Tracking and Fusion Group (CTFG) workshop, Ottawa, September 2012

Master's thesis, McMaster University, Hamilton, August 2012

Proc. SPIE Conference on Signal and Data Processing of Small Targets, Baltimore, Maryland, April 2012

  •  Prediction, tracking, and retrodiction for path-constrained targets 
Poster presentation, McMaster University, December 2011


  • Maritime simulator (C++)

The developed module in the Matlab is rewritten in C++ to generate target trajectories in large scale. Computational load is reduced by roughly 90 times with the new implementation.
Technologies: C++

  • Maritime simulator (Matlab)

A simulator generate target trajectories in maritime environment for different classes of ships. Port, hub, vessel type properties (speed, acceleration), birth rate, death probability, sojourn time for waiting in the port, loitering activities, great circle approximation are considered in this development.
Technologies: Matlab

  • Predictor plugin

A plugin to predict the locations of vessels considering the path-map, grid-map, landmass avoidance, and historical analysis of AIS data is developed.
Technologies: C++,  PostGIS, pgRouting, GDAL, OGR, pgAdmin, QGIS,  OSM

  • Adjoint sensitivity analysis in target tracking applications

Adjoint sensitivity analysis is applied to find out the optimum angular turn required for a target tracking scenario. This work is currently in progress.

  • Cognitive Radar Modeling exploiting Space Mapping Technology

An idea of modeling the cognitive radar based on space mapping (SM) technology is analysed in this work.

  • Performance Comparison of Non-linear tracking filters in Cognitive Radar and Traditional Active Radar

A case study to compare the information gain through use of the perception-action cycle mechanism (first step towards radar cognition) against Traditional Active Radar (TAR) for different nonlinear tracking filters. Following nonlinear tracking filters are considered: Cubature Kalman Filter (CKF), Extended Kalman Filter (EKF) and Unscented Kalman Filter (UKF).

  • Classification using SVM and EKF with RMLP

Performance of Support Vector Machine (SVM) vs. Extended Kalman Filter (EKF) with Recurrent Multilayer Perceptron (RMLP) is considered. 

  • Proximity Analysis Module

A proximity analysis module is developed in C/C++ which extracts the possible information from a given multi-target tracking scenario. Collision probability, collision detection, prediction and smoothing are considered in this module.

  • Path Prediction and Retrodiction with Real AIS Data

An AIS measurement data set, collected using a constellation of satellites, of vessels traveling in proximity to Indonesia  was used to test path prediction and retrodiction. The data set consists of thousands vessels over a span of three days, but with a varying scan rate. Since path map information is not available in the region, a path-map was generated with the available data set. A subset of data was selected and tested successfully with the algorithms developed in my master's research.

  • Destination Prediction for Under-Ground Mining

A framework was developed for evaluating the probability of reaching a safe destination for underground mine targets during a mine disaster such as explosion or entry/exit points blocked. Proposed framework was tested on a representative scenario to validate the theory.

  • Sensor Optimization for Multi-Target Tracking

A binary constraint optimization problem for selecting an optimal subset from large amount of sensors in a multi-target multi-sensor problem was considered. The binary relaxation approach was used to make the problem convex. Once the solution set for the relaxed problem is found using SEDUMI with CVX interface, a local search technique is used to find out the near optimum solution set for the original problem.

  • PCRLB Module for Multi-Target Multi-Sensor Problem

Posterior Cramer Rao Lower Bound (PCRLB)is a widely used performance measure in target tracking applications. In this project, a module was developed to calculate PCRLB in Matlab for a multi-target multi-sensor scenario without considering the information reduction factor.

  • Wireless Applications for PABX

This was my capstone project during my undergraduate degree. My group and I developed a functioning  integrated  telecommunication network for  Text, Voice, and Video communication with added features that enhanced the communication  in a PABX environment when compared  to the existing  facilities.