InnoLab: ITK Venture & Innovation Lab

Project goals

Nowadays, there are many fitness trackers and smartwatches that accompany us through the day and help us evaluate our daily workouts. Starting with simple parameters such as heart rate, step counts and training duration, fitness trackers often also record training intensity. Advanced devices are able to count the execution of certain movements specific to the sport and recognize them automatically in some cases. For this purpose, an Inertial Measurement Unit (IMU) is usually worn on the wrist in the form of a smartwatch and the activity is compared with taught-in patterns.

In InnoLab, motion analysis is to be extended by the use of smart textiles. The goal is motion recognition and assessment, i.e. automated identification of the movement and evaluation of correct or healthy execution. For this purpose, a prototype for a textile-integrated smart sensor network is designed and human motion is evaluated using Deep Learning.

ITIV participation

Within the InnoLab, ITIV investigates the execution of Deep Learning in embedded systems. This includes both the development of concepts for efficient hardware acceleration and the hardware/software codesign of neural networks in distributed systems. In addition, the methodology will be demonstrated using the example of motion analysis in smart textiles.