Sensor technology and AI for smart workwear to provide healthy and secure work conditions
 
Two pilots have been launched in the BIONIC research project. After a preliminary trial in July, the motion tracking systems are being put into test mode at Rolls Royce Power Systems in Friedrichshafen in October 2021. Last week, they have also been introduced at construction sites of Acciona in Spain. AI is used to evaluate the collected data in real time to facilitate detection of poor posture while working. Countermeasures can thus be taken at an early stage.
BIONIC stands for personalized body sensor networks with built-in intelligence for real-time risk assessment and coaching of workers, in all types of working and living environments. BIONIC systems are inconspicuously integrated into the usual work clothing and can monitor movements at the workplace during an entire shift. In case of unfavorable or health-endangering movements, the worker receives feedback. The systems can also be used during training and exercises.
Interactive Wear led the development of a modular and scalable sensor network to acquire and synchronize sensor data from distributed wearable sensors. Lightweight and robust textile cable were used for interconnection and invisible integration into work wear. Additionally, Interactive Wear has developed several economic variants with 2-4 sensors suitable for daily monitoring with a different application focus.
The concept of distributed sensor systems has already been successfully implemented in several projects. The UK based company Feraru Dynamics, for example, integrated a vibration-sensing system into a glove. This enables employees, who are exposed to (mechanical) vibration, to protect from such injuries as hand-arm vibration syndrome (HAVS). Equiltec, a German company, used the platform to develop a product that prevents neck discomfort and tension.
The systems can be used in multiple ways:

• For recording, analysis, and optimization of movements at a specific workplace. In this case, the systems consist of several motion and physiological sensors, usually 6-12, integrated in the typical work clothing.
• These systems could be also used for training programs with specific exercises.
• As a monitoring device for workers at the workplace, to measure the load over time from certain motions with high stress impact. For this purpose, systems with usually 2-4 sensors, are used.

As part of the i-compression research project, the Nursing Research Group of the Geriatrics Research Group at the Charité Universitätsmedizin Berlin has been using Interactive Wear’s MicroHub development system to record motion data, which were then classified by Interactive Wear. The motion patterns that needed to be detected were lying, sitting, standing, walking and climbing stairs, as well as two exercises typically used for mobilisation.

By using neural networks – convolutional neural networks with 3-5 levels implemented in Python/PyTorch – a hit rate of 85% to 93.2% was achieved. The same neural networks with Interactive Wear's Enhanced Features, on the other hand, yielded hit rates of over 99.5%. Since these hit rates can be achieved even with single-layer networks, the resource requirement for computation is also significantly reduced. A network created in this way can then run on a small, low-power ARM M0+ class microcontroller, in real time, with analyses performed every 2-3 seconds.

About sensor application development at Interactive Wear:

The Interactive Wear tools include hardware and software development systems that are suitable for a wide spectrum of body sensor network applications. The development systems and methods enable rapid prototyping of customer applications. The spectrum ranges from simple applications with 6 or 9-axis IMUs (inertial measurement units) to distributed systems with multiple IMUs and analogue sensors (e.g. pressure, strain) for sensor fusion as well as the integration of actuators. Data acquisition during the development phase is performed using a data logger that is adapted to the respective application and supports both live observations and analysis of recorded sensor data.

If you have any questions, please contact us at This email address is being protected from spambots. You need JavaScript enabled to view it..

i-compression is funded by the Federal Ministry of Education and Research (BMBF) over three years as part of the KMU-innovativ funding initiative for SMEs