2024 Global Student Innovation Challenge Rehabilitation Engineering and Assistive Technology (REAT) Taiwan

Congratulations to the following teams for their success in the Global Student Innovation Challenge:



Technology Category


GOLD - Speak for dysarthria (National Yang-Ming University, Department of Biomedical Engineering)
DVC 3.1
This study proposed a dysarthria voice conversion (DVC 3.1) system to help dysarthric patients have the ability to talk clearly and freely with people, which the voice conversion (VC) technology was used in this study. More specifically, we used a gated convolutional neural network (Gated CNN) with the phonetic posteriorgrams (PPGs) features to covert the features of dysarthric speech to target speaker speech features, firstly. Following, the WaveGlow vocoder technology was used to synthesize converted speech from this converted speech features. The results of speech intelligibility in the listening test and speech naturalness in MOS score show that our proposed system provided higher intelligibility and naturalness scores. These results suggested that the DVC 3.1 undoubtedly can be used as a communication assistive system for dysarthria patients to take advantage of in their daily live.


SILVER – NIRSafe (National Cheng Kung University, Department of Biomedical Engineering)
Near-infrared Spectroscopy in Sarcopenia during rehabilitation
The WTO predicts that by 2050, the global population of people aged 60 and over will increase from 900 million in 2015 to 2 billion. Several studies have shown that skeletal muscle produces and secretes molecules called myokines, which regulate brain function, including mood, learning, motor activity and neuronal injury protection. Furthermore, it may also lead to other brain-related diseases. Therefore, the purpose of this project is to develop a sarcopenia monitoring system that simultaneously evaluates muscle hemoglobin changes and brain oxygenated and deoxygenated hemodynamic changes using near-infrared spectroscopy to evaluate the effects of exercise programs. Sarcopenia is well documented to increase the risk of cognitive decline. Suppose we can quickly intervene and monitor brain changes in the early stage of sarcopenia, the therapist can change the exercise program in a rolling wave planning to increase the treatment effect and prevent brain lesions. For the device, we greatly reduce the size of the device, increase the signal quality, and achieve module expansion through circuit design. Finally, as a home rehabilitation device for patients with sarcopenia, this device can effectively rehabilitate and prevent brain lesions, so that the lives of the elderly and patients with sarcopenia can be more effectively improved.


BRONZE – W.J. (Chang Gung University, Department of Physical Therapy)
The smart kettlebell flips your life/ A portable physical assessment and exercise training kit
To popularize health promotion activities, we are committed to developing an easy-to-use and attractive product that can be used in homes, communities, and medical facilities. The user has to do is wear a head-mounted device connected to a smart kettlebell and follow the virtual therapist's instructions as if they were in real medical institutions or fitness centers. The analyzed assessment data can generate a health report and precise exercise prescription which helps to motivate them to regularly conduct self-assessment and exercise training. It will therefore advance the objective of early prevention for early intervention.


MERIT – CGU CSIE-OT (Chang Gung University, Department of Occupational Therapy / Department of Computer Science and Information Engineering)
Using deep learning-based pose estimation technology to develop a novel rehabilitation evaluation system
Target group: 1) Individuals with stroke experiencing arm and hand impairments. 2) Rehabilitation therapists.
Development Objective: To develop a reliable and objective AI and computer vision-based automated arm and hand rehabilitation assessment system for people with stroke.
Needs analysis: Currently, individuals with stroke are required to be evaluated by rehabilitation therapists during in-person sessions in the hospital. It is less accessible during the pandemic when person-to-person contacts are limited and for people living in remote areas. An accurate rehabilitation assessment system that can be done remotely is needed to address the above issues.
Prototype design and application: We developed an automated version of the ArmCAM test, an arm and hand assessment tool based on the AI skeleton posture detection and object detection technologies. We designed a toolbox that includes all the required assessment tools along with the instructions. To obtain the assessment results, patients only need to record their movements with any camera on their phone or computer following the instructions provided in the toolbox. By utilizing the system, users (e.g., therapists, people with stroke) can easily track changes after rehabilitation interventions without visiting the hospital.
Safety precautions: No safety precautions are required as the assessment system is non-invasive.


MERIT – DMolution (Chang Gung University, Department of Physical Therapy / Department of Computer Science and Information Engineering)
< FootHow >for diabetes foot care
The global diabetes population is increasing at a rate of 10 million per year. About 25% of diabetic patients will have diabetic foot, while 60% of the diabetic foot may have foot ulcers due to poor foot care or wound, and nearly 20% may lead to amputation. The most important thing to prevent diabetic foot is frequent screening and regular monitoring. At present, clinicians use manual methods to perform clinical examinations, including foot appearance, temperature, and sensory nerve neuropathy. The entire process takes about an hour and is quite time-consuming and labor-intensive, which leads to the current average diabetic patient being checked about every six months to a year. FootHow aims to solve the pain points of clinical diabetic foot ulcer and foot amputation, applying three core technologies (1) AI appearance inspection (2) infrared thermometer, and (3) intelligent sensory neuropathy assessment, which is original and novel and the leading technology in preventing diabetic foot ulcers in the world. The core value of FootHow is "Prevention, Precision, and Personalization" and the ultimate goal is to keep diabetic patients away from lower limb amputation.


Best Presentation – KMSTU (Kaohsiung Medical University, Department of Occupational Therapy / National Kaohsiung University of Science and Technology, Department of Mechanical Engineering)
Wearable Exoskeleton for Knuckle Muscle Rehabilitation
Stroke cases with hand spasm usually incapable of doing hand flexion and extension. In other words, stroke have an effect on hand function, which will fail to participate in activities of daily living. The research has shown that consecutive long-lasting and stretch-shortening cycle brings many benefit to stroke patients to rehabilitate their hand function, especially reducing stretch reflex sensitivity and muscle stiffness. In addition, repetitive movement can reduce muscle tension and stimulate the brain to learn the specific action.
We designed an exoskeleton to control hand movement. Patients can control the exoskeleton by themselves through the remote control which makes the device easier to use. The study and development stage have also shown that the three-finger function is the most important movement. Thus, we designed three-finger shape to train clients hand function.
Low cost and safety switch are two extraordinary advantages. The price between the existing electronic hand and our device produced by 3D printing technology vary greatly. Besides, the device equipped with a safety switch can reduce the usage of electricity.This exoskeleton device could use as a home-based rehabilitation way for stroke cases to train their hand function and thus achieve their independent living lifestyles.


Best Prototype – Exo Rehab (Chung Yuan Christian University, Department of Biomedical Engineering)
Customized Hand Exoskeleton Rehabilitation Device Many stroke patients suffer from upper extremity hemiplegia, and patients who have been injured for a long time after sports injuries can also cause muscle degeneration or atrophy and cannot coordinate movements. They need to go to the hospital for long-term physical therapy and rehabilitation by manpower or expensive but standardized machines. There are some problem that rehabilitation machines are expensive and cannot be applied to all rehabilitation patients on the market today. Therefore, this project designs a low-cost and customizable 3D printed hand exoskeleton. During rehabilitation, added the concept of "Mirror Therapy" that stimulates the brain with vision. In this research, a bending sensor is used to measure the bending angle of a normal finger, which is transmitted to the Arduino Nano microprocessor to convert into pulse width modulation (PWM), and then output to a motor to pull the 3D printed exoskeleton with nylon thread to drive the hand in need of rehabilitation.


Best Ergonomic – Liong Good (Jenteh Junior College of Medicine, Nursing and Management, Department of Rehabilitation Science)
Movement coordination trainer for trunk correction
This program amide at children with scoliosis, round shoulders, developmental coordination disorders, and mental retardation, such as poor posture, poor hand-eye coordination, and cognitive dysfunction. Scoliosis is usually congenital scoliosis of unknown origin, scoliosis of the thoracic spine, and acquired is caused by long-term poor posture or weight bearing. This product is from the shoulder to the pelvis of the back frame to improve the range of motion of the trunk, which can support the upper part of the body and the use of removable spheres to give the back muscles sensory stimulation, however the sense of stimulation of the human body's mechanical receptors,such as: skin, muscles, tendons to improvement muscle contraction to achieve postural stability.