Measuring and clustering moving objects

Gathering and processing data is essential as it concerns (road) traffic surveillance and management. Nevertheless, most of the current solutions concerning this are either insufficiently aligned with the corresponding real-life business processes, or are too expensive, or are not enough "interdisciplinary". We propose an approach that is claimed to be not only easy-to-implement but also not expensive to facilitate. The approach allows for clustering the "observed" moving vehicles and the data gathering is essentially based on a Forward Scattering (FS) radar. Such a solution offers a number of features such as: relatively simple hardware, an enhanced target radar cross section (compared to traditional radar), a long coherent interval of the receiving signal, robustness to stealth technology, and possible operation using non-cooperative transmitters. Further, a FS-based vehicle detection would not pollute the radio distribution with additional radio signals. Finally, the presence of GPS signals anywhere allows for a worldwide use. The data-analytics-related tasks (that are essentially driven by Machine Learning) are realized using tools, such as IBM SPSS Statistics and K-Means Cluster Analysis. The strengths of the proposed approach are partially justified by experimental data that is however only limited to the approach itself. Said otherwise, broader validation studies have not yet been conducted, such that the proposed approach is "compared" to other existing solutions. This is planned as future research. Still, the initial results obtained inspire us already to expect that the proposed signal processing of GPS signal shadows combined with Machine Learning can be successfully applied in practice for clustering moving objects in general, and in particular - for clustering vehicles in the context of road traffic.