Novel Nearest-Neighbor Data Association Algorithm with an Adaptive Threshold
Yuan Li, Dinghai Pan, Mengmeng Zhang, Zhi Liu · 2022 3rd International Conference on Big Data, Artificial Intelligence and Internet of Things Engineering (ICBAIE) · 2022
In the overlapping detection area of multiple millimeter-wave radars, the echo of a single target contains multiple radar measurement points. In the process of data fusion, these multiple homologous echo points need to be associated. In this paper, an adaptive threshold nearest neighbor data association (AT-NNDA) approach is proposed, which adds an adaptive association threshold judgment to the nearest neighbor algorithm, dedicated to solving the problem of multi-homolog echo point association. The proposed AT-NNDA is validated by collecting traffic scene data using smart vehicles equipped with multiple millimeter-wave radars. The experimental results show that, compared with the traditional foxed-threshold nearest-neighbor association algorithm, the AT-NNDA algorithm can effectively correlate multiple echo points, with a low false correlation rate and of high practical value.