Improving the Short-Range Perception of MIMO Radars with LO Feedthrough Topologies by Complex Sampling

In radar systems with local oscillator feedthrough topology, range shifts occur due to the time delays between the radar transceivers. These shifts can be corrected efficiently in postprocessing, except for the short-range spectrum, as disturbing peaks, caused by range shifts of the transmit-to-receive leakage, prevent the use of such system in short-range applications like automated parking. In this contribution, the capabilities of complex sampling, compared to real sampling, are analyzed both in theory and by measurements, with respect to the effects in systems with feedthrough topology. The imaging performance at very short ranges is investigated with an exemplary scenario, representing a typical automated parking situation. The measurements show that with complex sampling, the image peaks vanish and the noise level is reduced by 2.8 dB. In this way, radar images suitable for automated parking and other short range applications can be obtained.

I. Introduction

To achieve the level of perception needed for future advanced driver assistance systems (ADAS) or autonomous driving, the sensor data is fused to combine the object detections from different sensors [1]. Hereby, radar sensors are employed, as they are able to provide accurate information on range, velocity, and the direction-of-arrival (DoA) of surrounding objects, even under harsh environmental conditions [2] –[4]. For short-range scenarios, for example automated parking systems, the usage of radar data can improve the detection capability significantly, as the widely used ultrasonic sensors face problems in harsh conditions as heavy rain [5]. Therefore, radar systems with good detection capabilities, even at very short ranges below 1 m, are needed.