DEVELOPMENT OF VIRTUAL REALITY-BASED GAIT TRAINING SYSTEM FOR LOWER LIMB PROSTHETIC USER

• Main Author: Astari Pujiartati, Dwita
• Format: Dissertations
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/86688
• Institution: Institut Teknologi Bandung

For individuals with leg amputations, prosthetic legs are essential assistive devices that support their well-being. However, using prosthetic leg designs, especially for transfemoral amputees (TFAs), still presents various issues, one of which is gait disturbance. To improve gait, training that encourages user adaptation is required. There are two types of gait training: overground and treadmill. Treadmill training is more practical and safer, but it requires further design improvements to make the walking experience similar to overground and avoid monotony, which helps maintain training motivation. This study proposes integrating treadmill-based gait training with Virtual Reality (VR) technology, equipped with Kinect as a motion detection technology, to address these issues. VR is proposed due to its potential to provide varied and realistic environments, which is expected to enhance training motivation and gait performance. This system allows virtual environment stimuli to be processed by the human cognitive system to alter motor movements. In integrating VR into gait training, two main challenges arise: designing a VR- based training system and running gait training scenarios. Previous studies have shown a variety of VR-based gait training designs. A VR system design is needed that can measure gait performance, maintain VR sickness and fatigue within safe limits, and provide a high presence experience. Additionally, a task scenario design is needed to modify users' gait patterns. To address these challenges, this study undertakes two main stages: VR system design and task scenario development. The VR system design involves conceptual system design, followed by experiments to address specific needs that require further analysis. The conceptual design is based on an analysis of existing training systems, resulting in seven requirements. From the conceptual system design, five requirements still need to be validated: the ability to measure gait parameters, generate low VR sickness, maintain fatigue within safe limits, provide high presence, and support scenarios that can affect gait. Gait parameter measurement is tested by examining the influence of individual and motion detection technology variables on the resulting gait parameters. VR sickness and fatigue are analyzed by assessing the effects of individual variables and usage duration. Presence is analyzed by examining the effect of individual variables on presence factors. The ability to influence gait is evaluated through the effect of scenario variables, such as visual and audio stimuli, on gait parameters. The experiment involved 16 TFA participants and 16 non-amputee participants, all male. Data collection was conducted in two sessions. In the first session, participants walked for 70 seconds on a treadmill using a VR Head Mounted Display (HMD), and spatiotemporal parameters were measured using Kinect and Vicon motion capture. Data from non-amputee and TFA participants were compared to identify gait performance parameters. In the second session, participants walked on a treadmill for 40 minutes with 12 randomly ordered walking scenarios. During the experiment, gait was measured using Kinect and Vicon motion capture. Every 10 minutes, postural stability was measured with a force plate, VR sickness with the Virtual Reality Sickness Questionnaire (VRSQ), and fatigue with the Borg Rating of Perceived Exertion (RPE). At the end of the session, participants' experience was assessed using the Presence Questionnaire (PQ). VR sickness, fatigue, and postural stability data for non-amputee and TFA participants were compared at each measurement point to analyze the impact of duration and individual differences on the dependent variables. The task scenario data were processed using the data from the experiment. In this experiment, participants walked with 12 scenarios, involving visual and audio changes. The visual stimuli simulated walking on flat terrain, uphill, and on 10° inclined or declined slopes to accustom prosthetic users to walking on uneven ground. The audio stimuli simulated conditions with no audio, with audio in sync with step timing, and with audio slowed by 15%. The audio training aimed to improve step symmetry. Each condition lasted for 3 minutes, and the average of 20 gait cycles was compared to assess the effect of visual and audio stimuli scenarios on gait parameters. This study developed a VR system and task scenarios that influence users' gait. The resulting VR system can simulate 12 treadmill walking scenarios. The virtual environment is visually presented with a VR Head Mounted Display (HMD), integrated with Kinect motion capture as a motion sensor and equipped with a body harness for user safety. Guidelines for VR-based gait training system development and usage were also established: 1) Use cadence, step time, step length, stride length, and % opposite foot contact as performance parameters for TFA gait measurement; 2) Kinect can detect TFA disturbances in step time, stride length, and step length but is not as precise as Vicon, the gold standard, suggesting further development if practicality is prioritized; 3) Training duration significantly impacts VR sickness, fatigue, and postural stability, making duration an important consideration for TFA, with 30 minutes as the recommended limit; 4) TFA presence does not differ significantly from non-amputees, so no specific adjustments are needed, though adaptation and interface quality require attention as they scored lowest. Testing on gait scenarios revealed that visual stimuli can simulate uphill and downhill conditions, while audio stimuli may improve step symmetry with altered walking strategies, though repeated training sessions are needed. This study provides important contributions to promoting the adoption of new technologies in clinical practice to enhance the quality of life for amputee patients. The study serves as a guideline for developing VR-based gait training systems, including determining performance parameters, considering motion detection technology selection, setting safe training duration based on VR sickness and fatigue, and providing guidance on presence factors. It also offers alternative task scenarios proven effective in influencing TFA gait patterns.