10th Joint Seminar on Rehabilitation Engineering and Assistive Technology
Neural Electronic Engineering Laboratory Division of Biomedical Engineering for Health and Welfare Graduate School of Biomedical Engineering Tohoku University Sendai, Japan
Biocybernetics Laboratory Department of Biomedical Engineering Institut Teknologi Sepuluh Nopember (ITS) Surabaya, Indonesia
Schedule: Date: December 4, 2024 Time: 14.30 Surabaya /16.30 Sendai
Link: https://its-ac-id.zoom.us/j/99409063845?pwd=tCTzuOz9ptBXMR3arOOeSAFubhhFaF.1
Meeting ID: 994 0906 3845 Passcode: 811778
Presentation list: (1) Sensor and Computer Vision Based Wheelchair Navigation System for Detecting Obstacle and Tactile Paving Fathin Hanum Al’alimah1, Achmad Arifin1, Norma Hermawan1, Andra Risciawan2, Muhammad Lukman Hakim3, Nabila Alya Rahma1 1. Biomedical Engineering Department, ITS, Surabaya, Indonesia 2. Manufaktur Robot Industri (MRI), Surabaya, Indonesia 3. Industrial Mechanical Engineering Department, ITS, Surabaya, Indonesia
(2) A Basic Study on Pre-Training of Feedback Error Learning-Based FES Controller With Musculoskeletal Model: A Test on Elbow Joint Ryusuke Yokoya1, Takashi Watanabe2 1. Graduate School of Engineering, Tohoku University, Sendai, Japan 2. Graduate School of Biomedical Engineering, Tohoku University, Sendai, Japan
(3) Wearable Functional Electrical Stimulation (FES) System for Hip-Knee Joint Movement Restoration: ~Experiment with Normal Subjects~ Ni Luh Putu Nadila Aprilia Eka Natasa Putri, Achmad Arifin, Fauzan Arrofiqi Biomedical Engineering Department, Institut Teknologi Sepuluh Nopember, Surabaya, Indonesia
(4) A Feasibility Study of Stimulus Timing Determination Using Acceleration Information for FES Cycling Ryunosuke Sato1, Takashi Watanabe2 1 Graduate School of Engineering, Tohoku University, Sendai, Japan 2 Graduate School of Biomedical Engineering, Tohoku University, Sendai, Japan
Presentation Guideline: Presentation duration: 15 minutes Discussions: 5 minutes
Sensor and Computer Vision Based Wheelchair Navigation System for Detecting Obstacle and Tactile Paving
Fathin Hanum Al’alimah1, Achmad Arifin1, Norma Hermawan1, Andra Risciawan2, Muhammad Lukman Hakim3, Nabila Alya Rahma1
1 Biomedical Engineering Department, ITS, Surabaya, Indonesia 2 Manufaktur Robot Industri (MRI), Surabaya, Indonesia 3 Industrial Mechanical Engineering Department, ITS, Surabaya, Indonesia
Abstract
Individuals with visual impairments often depend on white canes to identify tactile paths and obstacles around them. For those who also use a wheelchair, however, managing both the cane and wheelchair presents notable physical challenges. This research introduces a wheelchair navigation system incorporating sensors and computer vision, designed to aid in detecting obstacles and following tactile paving paths. The system employs HC-SR04 ultrasonic sensors to measure distances to obstacles, ensuring safer and more informed maneuvering. Using OpenCV and TensorFlow Lite, the computer vision module achieved an object detection accuracy of 94.86% mAP50 across five object classes. Test results revealed that relying solely on a white cane produced the longest completion time, averaging 104.86 seconds, with the highest number of navigation errors totaling eight). In this scenario, the NASA-TLX workload score was 79.16, indicating high mental and physical strain. Comparatively, the proposed system demonstrated significant improvements, reducing the completion time to 44.65 seconds, nearly equivalent to the 42.38 seconds recorded under normal conditions. Although four navigation errors were observed, the system’s NASA-TLX score of 48.89 suggested a relatively low workload. When combining the white cane with the navigation system, the completion time averaged 70.12 seconds with six navigation errors, and the NASA-TLX score increased to 70, representing a moderate workload. These findings suggest that the designed system effectively minimizes both navigation time and user effort for visually impaired wheelchair users. Future development will aim to enhance system processing speed and reliability through parallel task execution, further optimizing its usability and performance in complex environments.
A Basic Study on Pre-Training of Feedback Error Learning-Based FES Controller with Musculoskeletal Model: A Test on Elbow Joint
Ryusuke Yokoya1, Takashi Watanabe2 1 Graduate School of Engineering, Tohoku University, Sendai, Japan 2 Graduate School of Biomedical Engineering, Tohoku University, Sendai, Japan
Functional Electrical Stimulation (FES) is a useful therapy for assisting and restoring purposeful movements in paralyzed limbs. The main methods of FES control are feedforward (FF) control and feedback (FB) control. A Feedback Error Learning (FEL) controller combines FF and FB controls, allowing the advantages of both methods to be utilized. However, the FEL-FES controller requires repeated FES applications for training the artificial neural network (ANN) used in FF control, which places a significant burden on the user. This study aims to develop an FEL-FES controller that does not require prior measurements from the user, thereby reducing the user’s burden during the ANN training phase. The proposed method reduces the need for user-specific measurements by training the ANN using a standard musculoskeletal model that does not depend on the individual user. In this process, the effects of differences in muscle characteristics between those used during training and those during control have not been examined, representing a challenge for the proposed method. To tackle this, this study focused on the nonlinear characteristics of muscles, created multiple musculoskeletal models with different nonlinear characteristics, and pre-trained the ANN using each model. The pre-trained ANN was then used in FEL-FES control with a musculoskeletal model that was not used during training and compared to an FEL-FES controller without pre-training. The test was performed though model simulation of elbow joint movement in the sagittal lane using two muscles, the biceps and triceps. The results showed that using a pre-trained ANN from a model with different muscle characteristics improved control performance compared to not using pre-training. Additionally, retraining the pre-trained ANN reduced the necessary learning process and enhanced control performance with fewer training iterations. Future work will explore whether ANN pre-training is also beneficial for differences in muscle characteristics beyond nonlinear characteristics.
Wearable Functional Electrical Stimulation (FES) System for Hip-Knee Joint Movement Restoration: ~Experiment with Normal Subjects~ Shoulder Joint Movement
Ni Luh Putu Nadila Aprilia Eka Natasa Putri, Achmad Arifin, Fauzan Arrofiqi Biomedical Engineering Department, Institut Teknologi Sepuluh Nopember, Surabaya, Indonesia
Stroke is the second leading cause of death and the third leading cause of disability worldwide, with an increase of 10.9 per thousand population in Indonesia in 2018. Hemiplegia due to stroke results in asymmetrical and unstable gait disturbances. Rehabilitation is crucial to help hemiplegic patients regain their walking ability. Previous studies have shown that electrical stimulation is effective for recovering lower body mobility but have focused only on the hip and ankle joints. This study examined an Functional Electrical Stimulation (FES) rehabilitation system to restore knee and hip joint moevements of swing phase, and tested in normal subjects simulated himeplegic gait. Burst durations of stimulation for as Iliopsoas, Rectus Femoris, Vastus Lateralis, Bicep Femoris Long Head, and Bicep Femoris Short Head were regulated based on cycle-to-cycle control method realized on MISO Fuzzy Logic Controller. Experimental test showed the embedded FLC regulated the hip and the knee movements, with a Maximum Hip Flexion in the Swing Phase (MHFsw) of 31.04 ± 4.65°, Hip Flexion at Initial Contact (HIC) of 28.38 ± 4.65, Maximum Knee extension in the Swing Phase (MKEsw) of 5.12 ± 2.51°, and Maximum Knee flexion in the Swing Phase (MKFsw) of 67.28 ± 4.43°. Further development could involve using SMD components for a more compact multichannel electrical stimulation system, employing supportive footwear for better gait detection in hemiplegic subjects, and integrating continuous stimulation to mitigate muscle spasticity.
A Feasibility Study of Stimulus Timing Determination Using Acceleration Information for FES Cycling
Ryunosuke Sato1, Takashi Watanabe2 1 Graduate School of Engineering, Tohoku University, Sendai, Japan 2 Graduate School of Biomedical Engineering, Tohoku University, Sendai, Japan
Hemiplegia after a stroke often causes muscle weakness and asymmetry in muscle activity, making basic movements like walking and standing difficult due to reduced balance ability. This study proposes a rehabilitation method combining a pedaling wheelchair and functional electrical stimulation (FES) to achieve symmetrical muscle activity by stimulating the paralyzed leg based on the muscle activity of the unaffected leg. However, electromyography (EMG), commonly used to measure muscle activity, is impractical for daily use due to setup difficulty and vulnerability to noise. In this study, a method to estimate timing of start and end of muscle activity of the vastus lateralis, the muscle responsible for knee extension, using forward acceleration signals obtained from inertial measurement units (IMU) attached to the wheelchair was examined. Experiments were conducted with two healthy participants under five conditions: cycling at 3 different speeds (slow, moderate, fast), changing seating position (change_position), and simulated hemiplegic cycling (hemiplegic). The relationships between the cycling speeds and the forward accelerations at the start and the end of muscle activity were approximated by the linear regression line. The results showed that the start points could be estimated by the approximated line with consistent accuracy regardless of cycling speed or individual differences. However, the end points exhibited increased errors due to individual differences. Future work will focus on these individual differences for end points and testing the method under various conditions, such as acceleration and deceleration during cycling.
10th Joint Seminar on Rehabilitation Engineering and Assistive Technology Neural Electronic Engineering Laboratory Division of Biomedical Engineering for Health
Kuliah tamu adalah sarana pengembangan wawasan serta pengetahuan bagi mahasiswa, sehingga mahasiswa mempunyai nilai tambah dalam pembelajaran serta pengalaman
9th Joint Seminar on Rehabilitation Engineering and Assistive Technology Neural Electronic Engineering Laboratory Division of Biomedical Engineering for Health