Time and Frequency Division Multiplexing in High-Resolution Imaging Radars with LO Feedthrough Topology

Abstract:
For high-resolution imaging radar sensors, typically a high number of transmit channels is required. However, with standard approaches such as frequency modulated continuous wave modulation and time division multiplexing (TDM), the maximum unambiguous velocity is significantly reduced due to the resulting long ramp repetition times. In this work, pseudo-random TDM combined with compressed sensing as well as frequency division multiplexing in the range domain are discussed to enhance high-resolution multiple-input multiple-output radar sensors with a large number of channels. While the former has previously only been validated and shown for systems with few transmit channels, a simple hardware implementation utilizing time delays in the local oscillator signal feed is presented for the latter. For both methods, the maximum unambiguous velocity can be increased by factors of 36 and 6, respectively. Due to the high number of transmitters, artifacts remain in the range-Doppler spectrum for both approaches, which are less significant for the reconstructed TDM. Therefore, this approach is used for measurements of an automotive scenario, where a pedestrian and a car at similar ranges but at different velocities (−1.5 m/s and 11.5 m/s) can be separated and estimated correctly.