IR and ToF measurement and algorithms for automatic signal evaluation for a mobile vehicle classification system (RaIT vehicle classification)

In order to collect data as precisely and cost-effectively as possible, technical measuring devices are generally used for this purpose, which today are only supplemented by manual recording in exceptional cases. Of particular interest is a mobile and non-invasive measuring technique that can be used at different locations as required, e.g. to record and evaluate the effect of guiding measures.

A neural network is to be used and calibrated accordingly to bring together the information from the technologies under investigation. It is to be examined whether it makes sense to condense raw data at the level of the individual measuring systems - radar, infrared and time of flight - or whether the neural network should be fed with the entire raw data. Once these questions have been clarified, the development of a functional model is planned, which will be optimised in terms of classification accuracy and from the point of view of energy efficiency.