Muhammad Hisyam Rosle
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Muhammad Hisyam Rosle has a total of 27 co-authors. Here are the top co-authors by number of shared publications:
- Abdul Nasir
Universiti Malaysia Pahang Al-Sultan Abdullah
4 shared publications - Devin Babu
Faculty of Electrical and Electronics Engineering Technology, Universiti Malaysia Pahang Al-Sultan Abdullah
3 shared publications - Shinichi Hirai
Department of Robotics, Ritsumeikan University, Kusatsu 525-8577, Shiga, Japan
3 shared publications - Zhongkui Wang
Ritsumeikan University
3 shared publications A. S. Jamaludin
University Malaysia Pahang
1 shared publication- Abbas Sufian Akmal Irfan
Malaysia-Japan International Institute of Technology, Universiti Teknologi Malaysia, 54100 Kuala Lumpur, Malaysia
1 shared publication
Latest Publications
- Open Access
- 28 August 2026
- Published by IOS Press
- p. 402-411
Predictive Maintenance for Heavy Diesel Machinery Using Machine Learning Model and Debouncing Finite State Machines
Unexpected equipment failure in heavy diesel engines leads to prohibitive operational costs and critical downtime during essential seasonal windows. This paper presents an integrated predictive maintenance (PdM) framework that combines a 6-state Finite State Machine for real-time monitoring with a multi-model machine learning ensemble. The architecture utilizes Random Forest for fault classification, Isolation Forest for vibration anomaly detection, and Linear Regression for oil pressure forecasting. Key design innovations include a temporal debouncing mechanism to filter transient noise, standardized Diagnostic Trouble Codes for technician interpretation, and a priority-based maintenance scheduling algorithm that weighs failure severity and asset criticality against estimated remaining useful life. The proposed framework was evaluated using three public benchmark datasets: Scania APS, NASA C-MAPSS, and CWRU Bearing. The system achieved accuracies of 99.3% on Scania APS, 96.2% on NASA C-MAPSS, and 98.4% on CWRU Bearing. On the Scania APS dataset, which serves as the closest operational proxy to agricultural diesel hardware, the model maintained a 0.2% false positive rate. This research contributes a systems-integration methodology demonstrating how heterogeneous AI techniques can be synergized for specialized machinery. While validated using public benchmarks, future work will focus on deploying this framework using live telemetry from engine hardware in real-world environments.
- 28 August 2026
- Published by IOS Press
- p. 402-411
Open AccessPredictive Maintenance for Heavy Diesel Machinery Using Machine Learning Model and Debouncing Finite State Machines
Unexpected equipment failure in heavy diesel engines leads to prohibitive operational costs and critical downtime during essential seasonal windows. This paper presents an integrated predictive maintenance (PdM) framework that combines a 6-state Finite State Machine for real-time monitoring with a multi-model machine learning ensemble. The architecture utilizes Random Forest for fault classification, Isolation Forest for vibration anomaly detection, and Linear Regression for oil pressure forecasting. Key design innovations include a temporal debouncing mechanism to filter transient noise, standardized Diagnostic Trouble Codes for technician interpretation, and a priority-based maintenance scheduling algorithm that weighs failure severity and asset criticality against estimated remaining useful life. The proposed framework was evaluated using three public benchmark datasets: Scania APS, NASA C-MAPSS, and CWRU Bearing. The system achieved accuracies of 99.3% on Scania APS, 96.2% on NASA C-MAPSS, and 98.4% on CWRU Bearing. On the Scania APS dataset, which serves as the closest operational proxy to agricultural diesel hardware, the model maintained a 0.2% false positive rate. This research contributes a systems-integration methodology demonstrating how heterogeneous AI techniques can be synergized for specialized machinery. While validated using public benchmarks, future work will focus on deploying this framework using live telemetry from engine hardware in real-world environments.
- Open Access
- 28 August 2026
- Published by IOS Press
- p. 24-33
Lightweight Vision-Language Model for Indoor Scene Monitoring in Healthcare Environment
This paper presents a lightweight Vision–Language Model (VLM) designed to support real-time indoor healthcare applications under resource-constrained deployment conditions. The proposed architecture integrates a ResNet-18 visual encoder and a BERT-base (uncased) text encoder to learn a compact multimodal representation for binary image–text matching, where a pair is classified as match or mismatch. A dataset of 600 image–caption pairs was constructed from open-source web repositories using four caption templates describing common indoor healthcare-like scenes and spatial relations. Positive and negative pairings were generated and split into 480 training and 120 validation samples. To assess generalization beyond the training distribution, an external test set was created using RGB images captured from a depth camera, paired with slightly rephrased prompts to evaluate robustness to linguistic variation. Experimental results on the test set demonstrate an overall accuracy of 84%, with a precision of 95%, a recall of 72%, and an F1-score of 82%, indicating reliable matching performance. In addition, an explanation-oriented trustworthiness evaluation was conducted using two questions, showing stronger performance for spatial-relation reasoning than for explicit human presence verification, reflecting the model’s scene-level matching formulation rather than object-level grounding. Overall, the proposed lightweight VLM provides an efficient baseline for context-aware indoor healthcare monitoring and motivates future work on integrating lightweight grounding modules to improve object-centric reliability in safety-relevant surveillance scenarios.
- 28 August 2026
- Published by IOS Press
- p. 24-33
Open AccessLightweight Vision-Language Model for Indoor Scene Monitoring in Healthcare Environment
This paper presents a lightweight Vision–Language Model (VLM) designed to support real-time indoor healthcare applications under resource-constrained deployment conditions. The proposed architecture integrates a ResNet-18 visual encoder and a BERT-base (uncased) text encoder to learn a compact multimodal representation for binary image–text matching, where a pair is classified as match or mismatch. A dataset of 600 image–caption pairs was constructed from open-source web repositories using four caption templates describing common indoor healthcare-like scenes and spatial relations. Positive and negative pairings were generated and split into 480 training and 120 validation samples. To assess generalization beyond the training distribution, an external test set was created using RGB images captured from a depth camera, paired with slightly rephrased prompts to evaluate robustness to linguistic variation. Experimental results on the test set demonstrate an overall accuracy of 84%, with a precision of 95%, a recall of 72%, and an F1-score of 82%, indicating reliable matching performance. In addition, an explanation-oriented trustworthiness evaluation was conducted using two questions, showing stronger performance for spatial-relation reasoning than for explicit human presence verification, reflecting the model’s scene-level matching formulation rather than object-level grounding. Overall, the proposed lightweight VLM provides an efficient baseline for context-aware indoor healthcare monitoring and motivates future work on integrating lightweight grounding modules to improve object-centric reliability in safety-relevant surveillance scenarios.
- Open Access
- 20 August 2026
- Published by Elsevier in SSRN Electronic Journal
A Skeleton-Grounded Vision–Language Framework for Biomechanical Critique of Amateur Tennis Strokes
Markerless human pose estimation provides accessible joint-level kinematic information, but its coordinate-based output does not directly support interpretable biomechanical feedback. Vision–language models (VLMs) can generate natural-language critiques, yet their application to raw sports images is limited by weak joint-level grounding and hallucinated or sycophantic responses. This study presents a two-stage skeleton-grounded vision–language framework for analysing amateur tennis strokes. First, a lightweight pose estimator detects COCO-17 keypoints and confidence scores, which are converted into a kinematics-aware visual prompt containing labelled joints, on-image joint angles, and confidence-modulated skeletal edges. Second, a 3-billion-parameter instruction-tuned VLM generates biomechanical critiques using structured prompting, with optional deterministic pose evidence, overlay-versus-raw visual contrastive decoding, and a motion-history skeleton composite for temporal cues. The framework was evaluated on 36 amateur tennis clips from the THETIS dataset, comprising 108 scored frames per condition, using a reference-free protocol based on a GPT-4o judge, hallucination-rate analysis, biomechanics-vocabulary coverage, and paired statistical testing. Compared with raw frames, the full skeletal overlay significantly improved biomechanical accuracy and completeness. Adding deterministic evidence or visual contrastive decoding produced the largest gains, increasing biomechanical accuracy from 2.56 to 3.09 and 3.00, respectively, while reducing the hallucination rate from 0.70 to 0.57 and 0.59. Temporal representations increased motion-related vocabulary but weakened grounding, whereas contrastive decoding restored grounding while suppressing some motion language. Overall, the results demonstrate that explicit pose-based visual grounding and decoder-level control can improve the reliability of VLM-generated sports feedback, although absolute critique quality remains moderate and further validation with expert-annotated references is required.
- 20 August 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessA Skeleton-Grounded Vision–Language Framework for Biomechanical Critique of Amateur Tennis Strokes
Markerless human pose estimation provides accessible joint-level kinematic information, but its coordinate-based output does not directly support interpretable biomechanical feedback. Vision–language models (VLMs) can generate natural-language critiques, yet their application to raw sports images is limited by weak joint-level grounding and hallucinated or sycophantic responses. This study presents a two-stage skeleton-grounded vision–language framework for analysing amateur tennis strokes. First, a lightweight pose estimator detects COCO-17 keypoints and confidence scores, which are converted into a kinematics-aware visual prompt containing labelled joints, on-image joint angles, and confidence-modulated skeletal edges. Second, a 3-billion-parameter instruction-tuned VLM generates biomechanical critiques using structured prompting, with optional deterministic pose evidence, overlay-versus-raw visual contrastive decoding, and a motion-history skeleton composite for temporal cues. The framework was evaluated on 36 amateur tennis clips from the THETIS dataset, comprising 108 scored frames per condition, using a reference-free protocol based on a GPT-4o judge, hallucination-rate analysis, biomechanics-vocabulary coverage, and paired statistical testing. Compared with raw frames, the full skeletal overlay significantly improved biomechanical accuracy and completeness. Adding deterministic evidence or visual contrastive decoding produced the largest gains, increasing biomechanical accuracy from 2.56 to 3.09 and 3.00, respectively, while reducing the hallucination rate from 0.70 to 0.57 and 0.59. Temporal representations increased motion-related vocabulary but weakened grounding, whereas contrastive decoding restored grounding while suppressing some motion language. Overall, the results demonstrate that explicit pose-based visual grounding and decoder-level control can improve the reliability of VLM-generated sports feedback, although absolute critique quality remains moderate and further validation with expert-annotated references is required.
- Open Access
- 1 August 2026
- Published by Institute of Advanced Engineering and Science in International Journal of Electrical and Computer Engineering (IJECE)
- Vol. 16 (4)
Bone strength analyzer and monitoring device for lower limb external fixation
The procedure for external fixator removal in lower limb fractures is typically based on radiographic data, subjects’ patients to ionising radiation, and provides minimal real-time information about the healing process. This project suggests a sensor-based Internet-of-Things apparatus, which will measure bone strength during the recuperating process based on load cells, HX711 amplifiers, and a Wemos ESP8266 microcontroller. The system provides real-time feedback in the form of LED indicators and a buzzer, whereas remote monitoring is supported by the Blynk dashboard. Measurement accuracy of over 90% was carried out as per the experimental validation conducted under the simulation of various loads (3-9 kg) and with clear stage indicators of critical, partial, and full recovery. The device offers continuous monitoring and objective bone healing evaluation with no radiation in comparison to conventional imaging. The limitation of the study is the limited range of loads in prototype testing, which could affect accuracy in particular situations. However, the results represent the possible clinical relevance of introducing real-time biomechanical surveillance into the process of fracture treatment, hence contributing to safer rehabilitation and more reasonable decisions related to the fixator removal.
- 1 August 2026
- Published by Institute of Advanced Engineering and Science in International Journal of Electrical and Computer Engineering (IJECE)
- Vol. 16 (4)
Open AccessBone strength analyzer and monitoring device for lower limb external fixation
The procedure for external fixator removal in lower limb fractures is typically based on radiographic data, subjects’ patients to ionising radiation, and provides minimal real-time information about the healing process. This project suggests a sensor-based Internet-of-Things apparatus, which will measure bone strength during the recuperating process based on load cells, HX711 amplifiers, and a Wemos ESP8266 microcontroller. The system provides real-time feedback in the form of LED indicators and a buzzer, whereas remote monitoring is supported by the Blynk dashboard. Measurement accuracy of over 90% was carried out as per the experimental validation conducted under the simulation of various loads (3-9 kg) and with clear stage indicators of critical, partial, and full recovery. The device offers continuous monitoring and objective bone healing evaluation with no radiation in comparison to conventional imaging. The limitation of the study is the limited range of loads in prototype testing, which could affect accuracy in particular situations. However, the results represent the possible clinical relevance of introducing real-time biomechanical surveillance into the process of fracture treatment, hence contributing to safer rehabilitation and more reasonable decisions related to the fixator removal.
- 18 April 2024
- Published by Springer Nature
- p. 83-93
Intelligent Machining Systems for Robotic End-Effectors: State-of-the-Art and Toward Future Directions
- 18 April 2024
- Published by Springer Nature
- p. 83-93
Intelligent Machining Systems for Robotic End-Effectors: State-of-the-Art and Toward Future Directions
- Open Access
- 1 January 2024
- Published by AIP Publishing in AIP Conference Proceedings
- Vol. 2898 (1)
Design of electroadhesion pad for improving grasping friction of soft robotic glove
A soft wearable robotic glove was developed to provide grasping assistance to patients with neuromuscular impairment. Sufficient friction forces is required in
- 1 January 2024
- Published by AIP Publishing in AIP Conference Proceedings
- Vol. 2898 (1)
Open AccessDesign of electroadhesion pad for improving grasping friction of soft robotic glove
A soft wearable robotic glove was developed to provide grasping assistance to patients with neuromuscular impairment. Sufficient friction forces is required in
- Open Access
- 1 November 2023
- Published by IOP Publishing in Journal of Physics: Conference Series
- Vol. 2641 (1) , 012027
Indoor Automated Fire Extinguisher System using Computer Vision
This paper presents the development of an automated fire extinguisher robot by employing fire recognition using computer vision. This project aims to develop a robot that can search, detect and extinguish small flames for indoor purposes. The robot is developed by applying open-source image processing to recognize the presence of fire. A control program was built to control the movement of the robot’s servo motor. The performance of the fire recognition was analysed in different light intensities and angles. The development of this project makes use of a camera and computer vision in place of various sensors, such as gas sensors, temperature sensors, and infrared sensors, to detect fires. Furthermore, a microprocessor is employed to operate the water pump and servo motor, which drive the nozzle to the position specified by the microcontroller. The results reveal that the proposed vision-based fire detection system has a high classification accuracy correspond to fire recognition.
- 1 November 2023
- Published by IOP Publishing in Journal of Physics: Conference Series
- Vol. 2641 (1) , 012027
Open AccessIndoor Automated Fire Extinguisher System using Computer Vision
This paper presents the development of an automated fire extinguisher robot by employing fire recognition using computer vision. This project aims to develop a robot that can search, detect and extinguish small flames for indoor purposes. The robot is developed by applying open-source image processing to recognize the presence of fire. A control program was built to control the movement of the robot’s servo motor. The performance of the fire recognition was analysed in different light intensities and angles. The development of this project makes use of a camera and computer vision in place of various sensors, such as gas sensors, temperature sensors, and infrared sensors, to detect fires. Furthermore, a microprocessor is employed to operate the water pump and servo motor, which drive the nozzle to the position specified by the microcontroller. The results reveal that the proposed vision-based fire detection system has a high classification accuracy correspond to fire recognition.
- 16 February 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Letters
- Vol. 6 (3) , 1-4
Soft Resistive Tactile Sensor Based on CNT-PDMS-Gel to Estimate Contact Force
Soft tactile sensors have been extensively studied due to their promising potential in estimating applied forces while exhibiting mechanical compliance. Such sensors have been fabricated from a variety of composite materials and various techniques to advance their applications. Currently, the development of soft tactile sensor based on carbon nanotubes (CNT)-polydimethylsiloxane (PDMS) gel has yet been investigated. This letter presents the design of a soft tactile sensor based on CNT-PDMS-gel composite that is sensitive to external forces. The sensor was composed of an elastomer, a rectangular conductive body made from a PDMS-CNT-gel composite, and a circuit board. The electrical resistance of the composite changed in response to the strain applied. Push and release tests were conducted, and the conductive body with 1% CNT was chosen due to high sensitivity, fast response, and stability in reading output. Experimental results, using feed-forward neural network, showed that the sensor can estimate the applied normal force based on the changes of resistance.
- 16 February 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Letters
- Vol. 6 (3) , 1-4
Soft Resistive Tactile Sensor Based on CNT-PDMS-Gel to Estimate Contact Force
Soft tactile sensors have been extensively studied due to their promising potential in estimating applied forces while exhibiting mechanical compliance. Such sensors have been fabricated from a variety of composite materials and various techniques to advance their applications. Currently, the development of soft tactile sensor based on carbon nanotubes (CNT)-polydimethylsiloxane (PDMS) gel has yet been investigated. This letter presents the design of a soft tactile sensor based on CNT-PDMS-gel composite that is sensitive to external forces. The sensor was composed of an elastomer, a rectangular conductive body made from a PDMS-CNT-gel composite, and a circuit board. The electrical resistance of the composite changed in response to the strain applied. Push and release tests were conducted, and the conductive body with 1% CNT was chosen due to high sensitivity, fast response, and stability in reading output. Experimental results, using feed-forward neural network, showed that the sensor can estimate the applied normal force based on the changes of resistance.
- 2 October 2021
- Published by Springer Nature
- p. 485-504
Holding, Grasping and Sensing of Prosthetic Robot Arm Like a Real Human Hand, a Journey Beyond Limits: An Extensive Review
- 2 October 2021
- Published by Springer Nature
- p. 485-504
Holding, Grasping and Sensing of Prosthetic Robot Arm Like a Real Human Hand, a Journey Beyond Limits: An Extensive Review
- 28 October 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 5 (1) , 159-166
Soft Tactile Fingertip to Estimate Orientation and the Contact State of Thin Rectangular Objects
- 28 October 2019
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 5 (1) , 159-166
Soft Tactile Fingertip to Estimate Orientation and the Contact State of Thin Rectangular Objects