Adaptive Tracking Strategy for the Positioning of Millimeter-Wave Radar Security Robots

Security robots often operate in environments characterized by low light and smoke, where millimeter-wave radar proves effective. However, the millimeter-wave radar’s point cloud is often sparse and noisy, potentially leading to positioning failure when employing point cloud matching. In this article, we propose a localization strategy for security robots based on millimeter-wave radar. The position of the security robot is deduced by clustering and tracking the sparse point cloud. In addition, radial velocity was used to design an adaptive tracking threshold, so that the targets in the two frames of data are within the tracking threshold regardless of the fast or slow motion speed of the security robot. The experimental results indicate that this method circumvents positioning failures associated with point cloud matching. In comparison to the radial velocity method, this approach enhances positioning accuracy by approximately 33.9%. Additionally, compared to the fixed tracking association threshold, this method exhibits a higher success rate in target tracking, leading to a 25.9% improvement in security robot positioning accuracy.

I. Introduction

Security robots play a pivotal role in ensuring the safety of environments by employing intelligent mobile technology, with their core functionalities dependent on environmental perception and navigation technology [1]. Of these, environmental perception is fundamental for the robot’s ability to accomplish various tasks autonomously [2]. Presently, environmental sensing relies predominantly on vision cameras and radar technologies, notably LiDAR [3], [4], [5], [6]. However, security robots often operate in challenging conditions such as smoke-filled, dimly lit, or dusty environments. In such scenarios, traditional sensors like vision cameras fail due to poor lighting, while LiDAR gets obstructed by smoke or dust, rendering them ineffective [7], [8], [9]. To overcome these limitations, the utilization of millimeter-wave radar emerges as a promising alternative. With advancements in automotive technology, mobile terminals, and robotics, microwave technology has witnessed increased research and application, approaching maturity [10], [11], [12]. This radar operates within a wavelength spectrum between centimeter and light waves, effectively combining the penetration capabilities of centimeter waves with superior spatial resolution, performing exceptionally well in adverse environments such as smoke and dimness [13], [14], [15]. Consequently, the utilization of millimeter-wave radar for environment sensing and localization of security robots has gained attention in recent studies [16], [17], [18]. In this article, we propose a novel millimeter-wave radar-based localization method for security robots. The main contributions of our solution are as follows. 1)

Given that cameras and LiDAR are unsuitable for dark and smoky environments, we propose the utilization of highly robust millimeter-wave radar for security robot localization.

2)

Since the point cloud of millimeter-wave radar is sparse, the point cloud matching method has the possibility of failure; limited by the accuracy of radial velocity estimation, the radial velocity method does not have high localization accuracy. We combine the advantages of these two methods to propose a new localization method.

3)

We use radial velocity to design an adaptive tracking threshold so that the targets in the two frames of data are within the tracking threshold regardless of the fast or slow motion speed of the security robot.