AIARA: Artificial Intelligence Enabled Highly Adaptive Robots for Aerospace Industry 4.0
The AIARA project seeks for novel solutions for flexible and versatile robotic manipulation using reinforcement learning methods. Conventionally, an agent is trained for a specific task of the robotic system. The agent should be retrained if the objective of the robotic task is changed. In AIARA, we aim at novel methods to develop an adaptive agent for a cluster of various tasks, such that retraining is avoided for certain environmental changes. The methods are then further applied to multi-arm collaboration scenarios to train multiple agents with inter-adaptation. Our ambition is to lead the way towards reliable manipulation of adaptive learning robots in complex environments.


Funding Sources (funded June 2020, 3 years)

  • NSERC CRD and CRIAQ in Canada
  • Lufo-8 in Germany
IMES: Integration of Artificial Intelligence into Manufacturing Execution Systems

The IMES project is intrinsically an intersection between Industrial Engineering, Robotics, and Computer Science that is aimed at achieving an intelligent adaptable manufacturing orchestration system, alleviating the obstacles of industry 4.0 implementation so that the manufacturers can cut operational costs, improve product reliability, and make the occupational environment safer. The project aims to enhance the efficiency and effectiveness of the planning, execution, and monitoring tasks in Manufacturing Execution Systems, by developing the Digital Twins of the production line comprising multi-agent systems powered by OPC-UA and proposing effective AI-based decision-support systems.


Funding Sources (funded October 2020, 3 years)

  • NSERC and IRAP in Canada
  • BMBF in Germany

IPES:  Intelligent Plant Execution Systems

Modern industrial plants similar to smart factories can take advantage of Industrial Internet of Things (IIOT) to interconnect different software agents, controllers and physical assets such as sensors and actuators. The design of such a complex interconnected system can be a challenging task. In this project, we are creating a digital twin for hydrogen-natural gas mixing plant to monitor and control processes, and predict and plan maintenance. A machine learning server consisting of multi-agents would be developed, where each agent analyses a particular aspect of each unit and makes the necessary decisions about the process control and maintenance planning. The Intelligent Plant Execution System is proposed as an add-on information analytics platform for plants with OPC UA compatible physical assets. With such a setting, the IPES can, by use of supervised and semi-supervised learning classification models, reliably handle uncertainties (e.g., leakage or equipment failure) that may arise in the plant without the need for halting the production/gas supply process or even waiting for human-intervention by exploiting the abilities of real-time high-fidelity simulations and digital twins.


  • Advanced Thermo-Fluidic Laboratory
  • ACIS Laboratory
  • Fortis
  • Hetek

 Funding Sources (funded January 2022, 3 years)

  • NSERC in Canada
  • FortisBC Engergy Inc. (FEI) in Canada
  • Hetek Solutions Inc. (Hetek) in Canada

AIEL-Photogram: AI-Enabled Aerial High-Precision Industrial Photogrammetry using UAV Formation

Photogrammetry is the process of creating three-dimensional models from 2-d images. This project focuses on high accuracy and high-fidelity photogrammetry model generation for images taken through a drone, to allow for condition assessment of structures and possibly geographical areas.


Funding Sources (funded June 2021, 3 years)

  • NRC

3D Active SLAM for Mobile Mapping of the Interior of Floating Roof Fuel Tanks

Manually scanning any types of environment is a labor-intensive task. This difficulty is magnified in crowded, confined, dark, or complex spaces. The same is the case with the floating lid tanks which are the target environment of this project. Currently, human mapping of the tanks requires technicians to undergo intensive training and then provide the manual labor to scan the interior. The goal of this project is to ease this burden by developing a robotic system capable of accomplishing the same task, autonomously.


Funding Sources (funded September 2019, 3 years)

  • MITACS in Canada
  • Rosen in Germany

Deep-learning for Distributed Intelligence Systems with Application in Robotics and Computer Vision

In this project, we will investigate shape-based or 3D computer vision for more accurate and robust object recognition and pose estimation by introducing unsupervised deep learning-based models and omit the need for labelling of new datasets. Moreover, we will investigate more complex robot operations that require flexibility and coordination in operation of adaptive robots and the vision system to automate manufacturing processes and quality inspection. Furthermore, we will develop a distributed intelligent system consisting of two robotic arms and multiple RGB-D cameras to examine and validate data-driven machine learning for practical implementation of adaptive robots in industry.  


Funding Sources (funded January 2020, 3 years)

  • MITACS and IRAP in Canada
  • BMBF in Germany
Responsive and Robust Object Detection for Industrial Point Cloud Applications

Our research objective is to create practical three dimensional or shape-based object detection methods that can support high precision industrial applications such as metrology and visual quality inspection. To achieve the required accuracies and computation performance, it is required that a combination of both the state-of-the-art machine learning methods and the classical statistical methods with their respective advantages are incorporated into a set of software solutions that can be optimally deal with different application scenarios.


Funding Sources (funded November 2021, 2 years)

  • NSERC in Canada
  • LlamaZoo in Canada

Machine Learning for process improvement and resource management in the oil and gas sector - Estimating oil and gas well life cycle using machine learning

Through research partnership with the British Columbia Oil and Gas Commission (BCOGC), we propose to develop and implement new algorithms that aim at predicting applications approval timelines and at forecasting future resources demand. The proposed project would provide a new reference for high efficiency service that could be implemented by members of the Western regulator forum as well as Oil and Gas companies across Canada. The research responds to a need to evaluate the new application process introduced by the commission, which allows operators to combine multiple activities (e.g. wells, pipelines, road…) in one application, rendering the estimation of timelines and resources harder to determine.   


  • UBCO Laboratory
  • ACIS Laboratory

Funding Sources (funded November 2021, 3 years)

  • NSERC in Canada
  • BC Oil and Gas Commission

Detection System for Screening of Household Hazardous Waste (HHW) in Recycling Facilities

Through research partnership with the Interior Freight & Bottle Depot (Interior Recycling), we propose to automate the sorting and classification of waste to reduce the burden of human labour necessary in traditional approaches. Our research objective is to develop a software that process image data of HHW containers to identify the product type and associated hazards.


Funding Sources (funded March 2022, 1 year)

  • NSERC in Canada
  • Interior Freight & Bottle Depot 

AI-Powered Management of Specialty Crops

In this project, we will propose novel AI-powered management schemes for the management of specialty crops to remove its dependence on intensive human labor. Yield estimation, anomoly detection, and autonomous navigation of multi-agent systems are some challenges of autonomous crop managment that will be investigated and solved using AI-powered methods, including machine learning, reinforcement learning, and learning-based control.


Funding Sources (funded December 2022, 1 year)

  • MITACS in Canada

Leading Indicator Generation from Cryptocurrency to Detect Fraud

This project aims at finding faster and more accurate fraud detection methods at the intersection of the Blockchain network and the MasterCard network (e.g., at the cryptocurrency exchanges). This, in turn, enables MasterCard to promptly notify its customers of whether a transaction appears fraudulent.


  • Mastercard

Funding Sources (funded Febraury 2023, 2 years)