Enhanced Target Tracking Based on Novel 5D Millimeter-wave Automotive MIMO Radar
Hengfeng Liu,Xiangrong Wang,Moeness G. Amin,Victor C. Chen
DOI: https://doi.org/10.1109/RADAR54928.2023.10371200
2023-01-01
Abstract:With the rapid advancement of radio frequency (RF) and the integrated circuit (IC) technology, millimeter wave (mmWave) multi-input multi-output (MIMO) radar, as one of the mainstream sensors, is widely used in automotive industry. The conventional 4-dimensional (4D) mmWave MIMO automotive radar can only estimate four parameters, those are distance, radial velocity, azimuth and elevation, which may render the vehicles insensitive to the targets with small radial movement. For example, vulnerable road users (VRUs), who suddenly cross the road, exhibit almost zero radial velocity for the driving cars towards them. Transversal velocity, which is a complement of radial velocity, is required to accurately track the moving target. In this regards, this work proposes a novel 5D mmWave MIMO radar, which is able to track the moving target in higher accuracy by adding the fifth dimension of transversal velocity compared with the conventional 4D counterparts. Different from our previous work, where the interferometry was performed at the receiver, this paper estimates the transversal velocity via combining receive beamforming and transmit interferometry. Finally, we test the proposed 5D MIMO radar in experiments, and the results manifest the effectiveness of the proposed novel 5D mmWave MIMO radar in terms of trajectory tracking.