University of Moratuwa crestCSE logoComputer Science & EngineeringUniversity of Moratuwa

intelliSense Lab

intelliSense Lab

Intellisense Lab, is a hub of cutting-edge research in integrated circuit engineering. It's a dynamic space where students dive into the world of IC engineering, fueling new discoveries. 
Students work closely with experienced mentors and lecturers at the university, gaining practical skills and insights that go beyond theoretical knowledge at the Intellisense Lab. The lab isn't just about learning – it's about becoming architects of innovation, ready to adapt in the ever-changing tech landscape.

Intellisense Lab is a launchpad for ideas that bridge imagination and technology. It's a place where innovation merges with practicality, and where students play a role in shaping how we interact with the digital world. Hence, it offers a collaborative environment for hands-on projects, from designing microprocessors to crafting efficient circuits and complex project ideas that take students to  world renowned competitions.
As students continue to learn, discover, and innovate, the lab remains at the forefront of pioneering technology.

Current key projects include:

  • Optimization of Deep Learning based Object Detection and Tracking Algorithms.
  • Autonomous Collision-Free Navigation of Unmanned Aerial Vehicles (UAVs) in Unknown Tunnel-Like Environments.
  • FabVis - an automated fabric quality inspection system(AHEAD funded and industry collaborated).
  • AgrIoT - a network of Soil Sensors and Drones for precision agriculture (NRC funded).
  • Xavier - a multi-robot system for 3D area exploration (industry collaborated).

 

Optimization of Deep Learning based Object Detection and Tracking Algorithms

Object detection algorithms are being used vastly in many industries. Applications based on robotics systems are getting more and more advanced day by day. Inorder to achieve higher precision and low power consumption these algorithms need to be optimized. Under this research project, a certain set of optimization methods will be implemented based on both computational efficiency improving and accuracy improving for real-world application. 

Research Objectives;

  1. Analyzing and identifying the structural architecture of the YOLO v7 detection algorithm and its contribution for the overall efficiency.

  2. Evaluating each layer's contribution to computational cost and developing methods to reduce computation cost while maintaining the base architecture.

  3. Comparison of efficiency gained by optimized architectures of other optimization methods.

  4. Evaluating the optimized algorithm by applying for real world application through prototype implementation

 

Autonomous Collision-Free Navigation of Unmanned Aerial Vehicles (UAVs) in Unknown Tunnel-Like Environments.

This research focuses on the development of an autonomous navigation system for Unmanned Aerial Vehicles (UAVs) to enable safe and efficient navigation through unknown tunnel-like environments. These environments present unique challenges, such as limited visibility, unpredictable obstacles, and the absence of GPS signals, making traditional navigation systems inadequate. The aim is to advance UAV capabilities for exploration, inspection, and surveillance in confined spaces, including mine shafts, stormwater drainage, and irrigation tunnels.

The methodology encompasses:

01. Sensor Selection and Integration: Utilizing sensors like LiDAR, cameras, and inertial sensors for accurate environment perception and mapping. 02. SLAM Algorithm Development: Adapting state-of-the-art Simultaneous Localization and Mapping (SLAM) algorithms for precise environment representation and UAV self-localization.
03. Path Planning and Navigation Algorithm Development: Creating algorithms for real-time, collision-free navigation that consider the environment's dynamic nature.
04. ML and AI Technique Integration: Implementing Machine Learning and Artificial Intelligence to enhance decision-making and adapt to changing tunnel conditions.
05. System Implementation and Evaluation: Testing the system through simulations and real-world experiments to assess navigation accuracy, collision avoidance, computational efficiency, and energy use.

 

The expected outcome is an autonomous navigation system that significantly enhances UAV operational capabilities in GPS-denied, confined environments, improving safety and efficiency for various applications.

 

Find out more on Youtube :  FABVIS on Jathika Mehewara (Newsfirst Sri Lanka) on 10.06.2022 (youtube.com)

                                             FABVIS Prototype Demonstration (youtube.com)


Director:
   Dr. Sulochana Sooriyaarachchi


Staff Members:
   Prof. Chandana Gamage
   Prof. Chathura de Silva