i Portfolio_Maaz Qureshi

Maaz Qureshi
Portfolio

AI/ML Research Engineer: Software Development

3+ years experienced R&D Software Development in ROS2 (Robot Operating System 2), C++, Python, and Artificial Intelligence frameworks (m23qures@uwaterloo.ca, LinkedIn).

Research Interest(s):
(a). Visual Intelligence, Autonomy i.e. Navigation (SLAM), and Motion Planning.
(b). Computer Vision and Robotics Perception (Radar, LiDAR, RGB-D) sensors.
(c). AI/ML: Imitation Learning/Reinforcenment Learning .

Jan, 2024 - April, 2025

Beyond the Visible

Advanced Volumetric Mapping (SLAM) with 5G-Connected Multi-Robots and 4D Radar Sensing

MASc. Thesis Research @ (a). ROBOHUB UWaterloo ; & (b). WSDL Labs
Supervisor: Dr. William Melek ; & Dr. George Shaker
Industry Partners: Rogers Inc

For smart factory(s) and warehouse(s) accurate and efficient indoor mapping and robotic inspections are vital for numerous applications, especially in large complex environments where traditional mapping methods often fail. We introduce an innovative method that addresses these challenges by combining advanced volumetric mapping with simultaneous localization and mapping (SLAM) techniques which utilize a visual camera, a 4D mmWave radar, and 5G technology. Our method improves the precision of visual sensors by allowing object(s) detection through occlusions and full obstructions, that is, walls and obstacles. The incorporation of 4D mmWave radar is crucial for detecting hidden structures, such as metal pipes behind walls, which are challenging for traditional sensors to detect. Using 5G technology, our solution ensures high bandwidth and efficient real-time data transmission, enabling the use of low-cost robots to perform sophisticated tasks by offloading computationally intensive work to a central server. Multiple experiments validate the robustness and effectiveness of the proposed method when deployed in real-world indoor mapping and inspection applications. -Github Repository: http://surl.li/xgfpxd

Fall 2024

CoBots
Panda Franka Emika

Algorithm development with motion planning framework using OMPL RRT Connect algorithm, integrating quaternion-based kinematics for autonomous high-precision measurements of Antenna-Radar. (Industry partner: Keysight Technologies)

Winter, 2024

Humanoid Robots
Human-Robot Interaction

Developed an (large language model) LLM and non-verbal communication framework for the NAO robot, enhancing collaborative task efficiency through gesture recognition for meal preparation in kitchen with humans (with ethics approval).

BASc Thesis Researh

Swarm Robotics
Multi-Robot Systems

Developed Swarm based custom Algorithm for multi agent UGVs to make pattern formation and navigation and extinguish the fire in the arena. 4 UGVs used for peer-to-peer synchronized topology via bluetooth for wireless data transfer.

January 2023 - December 2023

Underwater DAM Inspection
Field Robotics Ops

Digital twin based inspection of Neelum Jehlum (NJHEP) Dam HRT using ROV/AUV. Collaboration Hibbard Inshore (Michigan, USA) Saab Sabertooth double hull AUV via profiling and multi-beam sonar (advanced perception), and monochrome camera with attached tethered of 18 km. Project duration was 1 year, biggest underwater inspection project of South Asia

Fall, 2022

DAM Canal Inspection
Field Robotics Ops

Integrated the Oculus blueprint 750d sonar on the ROV (remotely operated vehicle). Captured the crack propagation till the concrete bed. Design and fabricated metal body frame for mounting sonar.Data acquisition for side beds, water velocity of 2.3 m/s, depth 9m.

Spring, 2022

Robotics in Health-Care
COVID-19

Developed and designed deep learning based facial recogintion and RFID-based autonomous pill dispenser nursing robot for path following. 6 different pills at the same time to the patient dispensing mechanism, also live stream to the doctor. The global application is targeted in hospitals to avoid contact with patients i.e. Covid-19.