MECHATRONICS focuses on the engineering of electrical, computer and mechanical systems. Intelligent systems may include Artificial Intelligence (AI) - based software systems, such as: expert systems, neural networks, fuzzy logic and genetic algorithms.  Within our research team the concept of Intelligent Mechatronic Systems is simply described as an electrical-mechanical device or system integrated with a microcomputer that has AI-based capacity to analyze sensor data and take action.  A functional prosthetic hand is an area of interest within the field of biomechatronics.

Prof Theo van Niekerk

Worldwide more than three million people live with upper limb amputations. Losing a hand drastically reduces a person’s quality of life due to the inability to perform everyday tasks. The Touch Hand (TH) has been part of South Africa’s first advanced mechatronics prosthetic hand research. The research has been in collaboration with Touch Prosthetics (www.touchprosthetic.com) and Prof Riaan Stopforth of the University of KwaZulu-Natal. The myoelectric prosthetic hand is required to give the amputee back the function of their hand. Prosthetic hands are commercially available, but at a high cost and these are usually procured from overseas sources with significant time delays. The need is growing for a functional low-cost prosthetic device produced locally that can be made available for amputees in developing countries. Figure below shown a CAD model of Touch Hand 4 with Static Digits in the Open and Closed position.

The research was performed by Kiran Setty, a Nelson Mandela University Mechatronics master’s degree graduate, who designed and developed the Touch Hand 4 (TH4) mechanical structure and sensor integration. He also conducted analysis and decoding of the sEMG signals to control the prosthetic hand.

Kyla Purdon, another Nelson Mandela University Mechatronics master’s degree graduate, included further feature extraction, the identification of signal thresholds of the sEMG signals, which the transradial amputee may be generating with actions they are trying to perform. An example of a low cost transradial myoelectric prosthesis is shown below. The software and embedded hardware (red box) was developed by Kyla Purdon, with a master’s entitled title:  Sensor Fusion Based Machine Learning Algorithm for Bio-Controlled Gripper.

Myoelectric prosthesis uses surface electromyography (sEMG) signals that are recorded from the muscles in the residual limb. A microprocessor device, which interprets the muscle signals and controls the motors via the signal that is pre-processed, perform the feature extraction.  These features serve as inputs to a machine learning algorithm to determine the control of the motors embedded into the Touch Hand. The sEMG sensors are placed on the transradial amputees’ partial limb in the strongest muscle groups. These signal thresholds become the input to the Support Vector Machine (SVM) Machine Learning (ML) Algorithm that classifies the output of the movements and hence drive the motors to rotate, open and close the hand. 

The research contributed in the development of a process with improved signal feature analysis and threshold detection. The improvement was a result of the process of finding movements and muscles that gave the best signals, although it might not respond to the movement that the prosthetic hand performed. Furthermore, the integrated machine-learning algorithm improved classification accuracy and overall response time.  The algorithm has the added function of being universal with only the thresholds needing to be updated through a calibration process.

The Advanced Engineering Design Group (AEDG, https://aedg.mandela.ac.za), led by Mr. Clive Hands, is currently involved in further structural and functional improvements of the existing myoelectric hand.  This is in preparation for participation in the 2020 Cybathlon prosthetic Olympics, which will be held in Switzerland. The recent establishment of the Nelson Mandela University Medical School and recent CoVid19 related responses across the university have pointedly exposed the opportunity for the Engineering School and the Medical School to collaborate in areas of common need to benefit local communities. 



Calado, A., Soares F. and Matos D., 2019, A Review on Commercially Available Anthropomorphic Myoelectric Prosthetic Hands, Pattern-Recognition-Based Microcontrollers and sEMG Sensors used for Prosthetic Control, Copyright IEEE.


  • Prof. van Niekerk works in the Department of Mechatronic Engineering and the Advanced Mechatronics Technology Centre (AMTC). The AMTC is an engagement unit within the Faculty of Engineering, the Built Environment and Technology with a focus on human capital development through education.  He provides academic leadership to the Nelson Mandela University - Siemens Certified Training Centre into Factory Automation and Drive Technologies.  Theo is a registered Professional Engineer and an active member of the Engineering Council of South Africa (ECSA) University Accreditation Team for Engineering Programmes. He holds a DTech in Electrical Engineering with a thesis entitled Monitoring and Diagnosis for Control of an Intelligent Machining Process. His research interest include factory automation, instrumentation and intelligent control systems. He is married to Eleanor and when time and body permits enjoys endurance sports.