mmPlace: Robust Place Recognition with Intermediate Frequency Signal of Low-cost Single-chip Millimeter Wave Radar

Abstract

Place recognition is crucial for tasks like loop-closure detection and re-localization. Single-chip millimeter wave radar (single-chip radar in short) emerges as a low-cost sensor option for place recognition, with the advantage of insensitivity to degraded visual environments. However, it encounters two challenges. Firstly, sparse point cloud from single-chip radar leads to poor performance when using current place recognition methods, which assume much denser data. Secondly, its performance significantly declines in scenarios involving rotational and lateral variations, due to limited overlap in its field of view (FOV). We propose mmPlace, a robust place recognition system to address these challenges. Specifically, mmPlace transforms intermediate frequency (IF) signal into range azimuth heatmap and employs a spatial encoder to extract features. Additionally, to improve the performance in scenarios involving rotational and lateral variations, mmPlace employs a rotating platform and concatenates heatmaps in a rotation cycle, effectively expanding the system's FOV. We evaluate mmPlace's performance on the milliSonic dataset, which is collected on the University of Science and Technology of China (USTC) campus, the city roads surrounding the campus, and an underground parking garage. The results demonstrate that mmPlace outperforms point cloud-based methods and achieves 87.37% recall@1 in scenarios involving rotational and lateral variations.

Figures (8)

• Figure 1: Our single-chip radar place recognition system. It finds the same location in a pre-built map database based on the given query data.
• Figure 2: The mmPlace system consists of three main components: (1) heatmap generation, (2) feature extraction, and (3) heatmap concatenation. The heatmap generation module performs range estimation and angle estimation on the IF signal to generate the range azimuth heatmap. The feature extraction module employs a spatial encoder on the heatmap to produce the place descriptor. The heatmap concatenation module employs a rotating platform and concatenates the heatmaps over a full rotation cycle.
• Figure 3: Visualization of the range azimuth heatmap. The heatmap shows the location of objects and indicates the power of the signal reflected back from them. As an example, the bright spots in the heatmap correspond to the light poles seen in the photo.
• Figure 4: Overview of the spatial encoder. After obtaining the range azimuth heatmap, the spatial encoder performs feature extraction on the heatmap to obtain the place descriptor.
• Figure 5: The radar is deployed on a rotating platform. The rotating platform rotates horizontally over 180 degrees at a speed of 150 degrees per second, which effectively captures data from multiple angles.