Eileen Su Lee Ming

Enhancing effectiveness of virtual reality rehabilitation system: Durian Runtuh

Stroke is a global health problem and leading cause of severe disabilities. Majority stroke survivors have impaired upper limb and body balance problems which influence their performance for activity of daily living. Kinect-based system for home-based virtual rehabilitation becoming popular and be proven much more effective than regular therapy. However, most of the Kinect Rehabilitation Systems are lacking in reported in functional and validation data because majority of systems are only at a proof concept stage. Besides that, lack of proper exercise routines to convert conventional rehabilitation training into virtual reality gaming training. Using research and recommendations from physiotherapists, Kinect-based virtual reality game-Durian Runtuh was developed to build an effective rehabilitation system. Durian Runtuh consists of four exercise routines to train patients body balance and limb impairment. It is a simple-to-use system with attractive game environment to increase the motivation and attention of patients. This paper will present the design of the virtual reality system-Durian Runtuh and enhancing effectiveness of the Kinect game.

Simultaneous localization and mapping survey based on filtering techniques

Simultaneous Localization and Mapping (SLAM) problem has been an active area of research in robotics for more than two decades. This paper reviews SLAM based on different filtering techniques used to do the state estimation of the mobile robot. The filtering techniques included in this study are Kalman filter, particle filter, H infinity filter. It can be concluded that each filtering technique has its own advantages and disadvantages as it is very dependent on the situations. Kalman filter is much suitable for dealing with Gaussian distribution. Particle filter is selected for large-scale environment as its computation complexity is logarithmic compared to Kalman filter which has quadratic complexity. H infinity filter is used to improve the convergence of SLAM system.

Investigation of upper limb movement during badminton smash

This paper discusses investigation of dynamic data of upper limb movement included wrist, elbow and shoulder during the badminton smashing performed by using wireless inertia sensor system. Research shows that the wrist movement of the player contributes a lot during the badminton smash compare to the elbow and shoulder movement. Finding of this project will help badminton players to come out with a training system or protocol that help to improve the effectiveness of training and enhance their performance.

iRest: Interactive rehabilitation and assessment tool

The ability to perform activities of daily living such as eating, drinking or writing is limited after a stroke. Rehabilitation is to encourage the recovery process and to define effective rehabilitation procedures, quantitative assessment should be included. Conventional assessment scales such as Fugl-Meyer and Motor Assessment Scale have several limitations as the scales are subjective, provide only rough estimates, and are highly dependent on the ability of the trained physiotherapist; making it difficult to quantify impairment and disability. Using these methods, small changes due to stroke are often undetected. Therefore, precise and objective measurement of motor impairment is needed to monitor and quantify the patients progress. This paper presents the design and development of an interactive rehabilitation and assessment tool, iRest. This system does not use any robotic actuator but only uses a small DC motor to generate vibration. Three robotic assessment modules were developed as a virtual reality (VR) environments namely Draw I, Draw Diamond and Draw Circle. The design of this VR considers the required movement in activities of daily living, including isolated and combined movements.

Shirt pattern recognizing and tracking behavior analysis using SURF on the moving human target

In the field of human following using mobile robot, human as the object for the detection is not the only requirement to be considered. Besides detection, continuous tracking and recognition is also another important criteria. In order to track and recognize a human, an effective algorithm needed to be implemented in the system. Speed Up Robust Feature (SURF) algorithm is a fast and robust algorithm for local, similarity invariant representation and comparison which mainly used for image analysis. In this paper we made the analysis using the SURF on the moving human target based on the shirts pattern and the results were shown in the discussion.

Development of InnovaBoard: An interactive balance board for balancing training and ankle rehabilitation

Ankle sprain is one of the most common injuries among sports with high occurrence and recurrence rate of over 70%. This injury can be prevented with proprioceptive trainings, such as wobble board training. These training can increase range of motion in the ankle joint complex and strengthen musculature of the foot ankle and lower leg. However, most basic wobble and rocker board lack biofeedback during training and objective assessment. Therefore, the aim of this project is to develop a portable and interactive balance board that provides multiple levels of training difficulty, combined with objective assessment and interactive virtual reality games. This paper presents the development of the InnovaBoard, its sensor performance evaluation and functionality in assessing balancing skill of healthy subjects. Results show that InnovaBoard has high accuracy of ±0.04% and repeatability of ±0.06 degrees. Preliminary trial shows the developed assessment program is able to differentiate significant difference (p = 0.05) between double-leg stance balance test with eyes open and close.

Vision Recognition on Line Pattern Following Behavior on POB-Bot

Robust development in the industries had lead to the fast development of the robots and sensors. Mobile robot is able to replace human power for many things due to its low cost and high accuracy. Types of navigation of the mobile robot had been in the research for past decades, the simplest and straight forward method of mobile robot navigation will be line following. Line following basically allows the robot to move along the predefined line. This paper is proposing a novel method of line following for more functions for mobile robot. 12 patterns have been defined to perform different functions. POB-Bot has been used to recognize these patterns using its POB-Eye. The result shows that the accuracy of the POB-Eye is promisingly high for simple pattern.

Analysis of Human Hand Kinematics: Reaching Upward Movement

Reaching movement is one of the important training for stroke patient to recover the impairment arm as this movement is critical for activities of daily living. A studied conducted to investigate the minimal requirement that a rehabilitation robot should have in order to train reaching movement in upward after the impairment arm start to overcome gravity. Result show some convex, linear and concave trajectory during reaching movement in upward direction depended on the level of the target distance. This study can be applied for designing a new simpler or improvement of the existing rehabilitation robot to train the upper limb for stroke patient.

Design and Development of an Omnidirectional Mobile Robot

Most mobile robots use differential-drive concept, where they are equipped with two actuators that permit only single-direction rotation at a time. This concept limits the robots navigation because its orientation must always change according to the direction of movement. This paper presents the development of an omnidirectional mobile robot that uses three actuators, aligned in 120 degrees separation and each attached to an omniwheel. By manipulating actuator speed, the robot can navigate to any direction without changing its orientation. Using NIsbRIO9632xt as the main controller, navigation algorithm is implemented in LabVIEW, integrated with PID controller to fine-tune robot movements.