Digitisation in the areas of Industry 4.0 and autonomous driving is rapidly gaining in importance. Wireless communication networks form the necessary basis for this. Since their performance cannot be measured at all place and at all times, it will be necessary to predict performance.
For this purpose, in the 5G-AI-MLab we determined which performance characteristics can be predicted how well on the basis of which data. We developed the fundamentals to estimate the quality of wireless communication (focus on 5G) both for points in time and for places where no measurements are made. The quality of the connections includes latency, data rate and reliability.
The main goal of the project was the integration of different information sources and the subsequent prediction based on this information. Examples of information sources are past measurements, reference measurements, network configuration parameters, current network information, base station positions and maps/sketches. For forecasting we used machine learning methods.
In general, in the 5G-AI-MLab we developed the predictions that will allow us to predict the performance of wireless (5G) networks in the future. Such predictions are especially important for many future applications, especially networked vehicles and Industry 4.0.