Electrical and Computer Engineering

Biomedical and Intelligent Systems

The biomedical and intelligent systems group consists of five faculty members from the Department, two faculty members cross-appointed from the School of Computing, one faculty member cross-appointed from the Department of Mechanical and Materials Engineering, and one faculty member cross-appointed from the Royal Military College. The group utilizes state-of-the-art robotic, imaging, and sensory systems as well as technological advancements in controls, system identification, mechatronics, computer vision, and signal and image processing to lead innovative research in the areas of i) intelligent and autonomous systems, ii) teleoperation and virtual environments with medical applications, and iii) biomedicine, biomechanics, and biorobotics. The group research activities are jointly conducted by members of the group or in collaboration with leading national and international researchers and commercial organizations.

Research in intelligent and autonomous systems focuses on modeling and control of discrete event systems (DES) for autonomous control of multi-agent systems, and computer vision for object recognition, pose determination, and tracking for intelligent robotic applications. Recent research on multi-agent discrete-event systems is unified by a common theme of integrating control, communication and knowledge. This research is in collaboration with globally-known leading researchers from the USA. Research in computer vision focuses on the development of a vision capability for the autonomous grasping of free-flying satellites in space. This globally recognized research is conducted in close collaboration with MacDonald Dettwiler Associates (MDA) Space Missions and the Canadian Space Agency. In addition, other applications of computer vision, robotics, and intelligent systems are actively pursued for robotic gaming systems, advanced human computer interfaces, high frame rate sensor development, and automated analysis of urban scenes.

Research in teleoperation and virtual environments involves control systems for master-slave telerobotic and haptic simulation systems with applications in medical training and diagnosis. Telerobotic systems project human sensing and manipulation ability to remote locations and virtual worlds using a force-feedback, or haptic, device that is interfaced with the user and a slave robot that performs a desired task on a remote environment. Two emerging medical applications of haptics and virtual environments have recently been pursued collaboratively with researchers from Queen's University, McMaster University, and Quanser Consulting Inc. in Markham, Ontario. Research in the development of a haptic-based ultrasound simulator will be used for training of ultrasound sonographers and diagnosis of patients in remote areas of Canada where access to expert radiologists is limited. Research in human haptic guidance control systems will be utilized in the development of a haptic-based surgical simulator that will be used for teaching minimally-invasive surgery manoeuvers and skills.

Research in biomedical engineering and computing includes treatment prediction, human motor and posture analysis, and control of prosthetic arms. The research in biomedicine concerns nonlinear system identification techniques applied to prediction of treatment response, clinical outcome, and cancer survival. This work has important significance both nationally and internationally, especially for microarray data analysis. In particular, this research was first to predict the treatment response of a group of acute myeloid leukemia patients with a new nonlinear processing technique, as featured in the prominent journal, Nature Reviews Drug Discovery, in 2002. Another collaboration with a leading researcher at MIT focuses on predicting clinical outcome for a large group of cardiac patients. The collaborative research in biomechanics involves the use of the electromyogram (EMG), which is a detectable signal associated with muscle contraction, and upper body acceleration for the study of human motor system and the analysis of body posture. The nonlinear processing technique mentioned above has been applied to EMG data recorded from lifting exercises in order to identify the load in the hands and on the lower back. In research supported by Defense Research and Development Canada, loading effects on human performance have been examined in several studies involving heavy backpack loads, and a biomechanical model of backpack load carriage has been developed. Research in biorobotics involves the nonlinear mapping between EMG and wrist force to design advanced controllers for prosthetic arms.

G. Fichtinger

  • Computer-Integrated Surgery, a fascinating and complex field that covers medical imaging, image computing, scientific visualization, surgical planning and navigation, robotics, biosensors and, perhaps most importantly, integration of all these into workable clinical systems and translating them to effective clinical use.
  • Robot-assisted minimally invasive percutaneous (through the skin) surgery performed under real-time image guidance, with primary application in the detection and treatment of cancer.
  • Laboratory for Percutaneous Surgery (Perk Lab)

M. Greenspan

  • Pose determination and object recognition in 3D range image data
  • Machine vision methods for pose estimation and tracking
  • Efficient robotic manipulator collision detection and motion planning, including parallel methods
  • Applications of machine perception to space robotics and remote teleoperation
  • Applied computational geometry
  • Robotics and Computer Vision Laboratory

K. Hashtrudi-Zaad

  • Control systems design and analysis of haptic and telerobotic systems
  • Medical applications of haptics and telerobotics
  • Robot-assisted needle insertion
  • Modelling of human arm biomechanics
  • Biosensory systems and signal processing
  • Environment impedance identification
  • Robot motion and contact control
  • Control and payload monitoring of heavy-duty hydraulic machines

M.J. Korenberg

  • Time series analysis and system identification
  • Nonlinear system representation.
  • Estimation of sinusoidal and exponential series models from short-duration nonuniformly spaced samples.
  • Protein classification and genome sequencing
  • High resolution Raman Spectroscopy

E.L. Morin (Group Coordinator)

  • Electromyogram signal processing :
    • Estimate and predict muscle and joint activity
    • Assessment of pathological conditions in joints
    • Develop control algorithms in upper limb prostheses.
  • Ergonomic data processing: Applied to human subjects carrying heavy loads
  • Ergonomics Research Group

K. Rudie

  • Decentralized control of discrete-event systems
  • Modeling of emergency response protocols
  • Automation of software development for concurrency control code
  • Application of discrete-event systems to telecommunication protocols
  • Use of formal reasoning about knowledge and modal logic
  • Discrete-Event Control Systems Lab

Parvin Mousavi (Computing)

Aboelmagd Noureldin (RMC)