Radar accuracy improvement by pattern multiplication for automotive radar systems and other sensing scenarios

The automotive radar market is growing, and several sensor systems are now present on a car, with computing power and increased complexity for the system architecture. While radar is continually researched and improved, evolving architectures often require advanced processing techniques. In this paper, angular radar target estimates are multiplied and investigated in the context of automotive radar for selected millimeter-wave system architectures. While previous radar systems preserved the template of one function per one microchip or radar module, improved performance can be achieved with a diversification of the adopted signal processing techniques across multiple chips. Also, by introducing multiplication of the angular target response whilst considering antenna beam steering, resolution and radar detection accuracy can be improved. Basically, the multiplication technique can offer reductions in the side-lobe level during radar target detection, made possible by collecting several data samples in the time-domain whilst processing them together and this achieves a better radar response. In this paper, this technique is mathematically outlined and then considered for two distinct radar system architectures: a multiple input-multiple output (MIMO) radar and a scanning radar using a Butler matrix transmit array, with a noticeable side-lobe level improvement of 13 dB improved radar resolution.