Off-Grid Radar Processing and Target Detection

Radar systems are increasingly used in the field of driver assistance systems and autonomous driving. The application areas are very different and therefore also the requirements to the radars. Typically, different optimized radar sensors are used due to different requirements and tasks (parking assistant, ACC, ...). Since these radars are designed for a specific purpose, they are not very flexible and cannot be used for others. New radar concepts such as OFDM offer a high degree of flexiblity concerning signal generation and evaluation. OFDM radars allow to change signal parameters during operation. Therefore, they are able to adapt to different situations in traffic. At the Institute of Microwave Engineering, we have been working on the realization of such OFDM radars for some time now. Within the scope of our research, several concepts for digital radar signal processing and evaluation have already been developed and tested using simulation and measurements.

In one of these new radar concepts as well as MIMO-OFDM operation (using multiple transmitters), a subsampled transmit signal is used which has a specific signal pattern. This leads to an incomplete receive signal that needs to be completed at the receiver before evaluation. For this purpose a specific signal reconstruction. However, digital signals, compared to continuous ones, cannot resolve values and oscillations arbitrarily accurately. So-called off-grid methods can be applied to counteract this issue and improve the quality of the signal post-processing and target detection. Therefore, in this work, algorithms from literature should be investigated, evaluated, implemented and adapted to the existing signal processing. Then, the performance of the algorithms should be evaluated using simulations and radar measurements.

Tasks

Literature study and research, algorithms implementation and simulation in Matlab, verification using measurements

Prerequisites

good knowledge of signal processing and experience with Matlab is advantageous

Misc

Start: as of now
The focus of the work can be elaborated according to one's own preference