Data-Driven Localization in Industry 4.0 and Logistics

Object localization is a fundamental requirement for many applications in logistics and Industry 4.0. Fraunhofer IIS has extensive expertise and many years of experience in the field of localization. Radio-based localization solutions are frequently used; these typically measure the propagation times of radio signals emitted by mobile objects. The accuracy of the localization system depends largely on the optimization of fusion and system parameters across diverse and sometimes changing environments.

Fraunhofer IIS is therefore working on methods for more robust position determination. With the help of machine learning, positions are to be calculated automatically by using reference measurements as empirical data. The raw data consists of measurements from a localization system, which are used to train deep neural networks that can replace parts or entire position calculation procedures.

This technological development is being further advanced within the framework of the ADA Lovelace Centers. To this end, topics such as fine-tuning in changing environments, active learning for model calibration, motion models, and hybrid sensor fusion are being addressed.