Performance Evaluation of an Indoor Positioning Scheme Based on Millimeter Wave Radar

With the widespread adoption of the Internet of Things (IoT), the importance of high-precision indoor coordinate information has become increasingly evident in both civilian and military applications. In certain scenarios,, traditional active wireless indoor positioning solutions are unable to resolve positioning issues when targets lack auxiliary devices, with positioning errors typically exceeding 1 meter. In recent years, to address this issue, passive positioning sensing technology has garnered significant attention. Among these, millimeter-wave radar stands out due to its ability to resist light interference and its strong penetration capabilities, demonstrating unique advantages in indoor positioning applications. However, with the development of society and radar technology, millimeter-wave radar positioning technology combined with linear regression models has gradually been applied to multiple industries, including the automotive and intelligent transportation sectors. This study tested the performance of millimeter-wave radar-based position estimation technology in indoor environments, using a linear regression model to compensate for the limitations of millimeter-wave radar positioning. Experimental results demonstrate that millimeter-wave radar can track the dynamic positions of one or two pedestrians along different paths: in straight-line motion scenarios, the standard deviation of positioning error can be controlled within 15 cm; in complex paths such as rectangular trajectories, $90 \%$ of positioning points have errors less than 50 cm. Compared to traditional wearable device positioning systems, the experimental results demonstrate higher accuracy and dynamic tracking performance in actual indoor environments. This paper combines machine learning technology with millimeter-wave radar to reduce the average error to 65 cm, significantly improving indoor positioning accuracy. The proposed method can reduce the positioning error of millimeterwave radar to within 50 cm, meeting the accuracy requirements of scenarios such as medical care and industrial monitoring.