The Rise of Radar for Autonomous Vehicles Signal processing solutions and future research directions

Automotive radar is the most promising and fastest-growing civilian application of radar technology. Vehicular radars provide the key enabling technology for the autonomous driving revolution that will have a dramatic impact on everyone's day-to-day lives. They play a central role in the autonomous sensing suit because of the significant progress in the radio-frequency (RF) CMOS technology that enables high-level radar-on-chip integration and thus reduces the automotive radar cost to the level of consumer mass production. However, this would not be sufficient without high spatial resolution performance, which can be obtained by multiple-input, multiple-output (MIMO) and cognitive approaches at a lower cost. The uniqueness of automotive radar scenarios mandates the formulation and derivation of new signal processing approaches beyond classical military radar concepts. The reformulation of vehicular radar tasks, along with new performance requirements, provides an opportunity to develop innovative signal processing methods. In this article, we first revise conventional techniques for signal processing in automotive radar. Then, we emphasize the limitations of the historically driven conventional processing approaches in practical roadway scenarios and present alternative signal processing solutions. Finally, we propose several future research directions to enhance vehicular radar performance.