Sensor Response-Based Localization Using Spatial Correlation of Nongeotagged Data

Abstract

Sensor devices in future wireless communication systems such as beyond fifth generation (Beyond 5G) and sixth generation (6G) require highly accurate location information. Owing to high power consumption and cost, global navigation satellite system receivers may not be installed on sensor devices, requiring the localization of sensor devices on a wireless sensor network (WSN). In many WSN applications, a spatially dense sensor installation is required to achieve satisfactory coverage. When object detectiontype sensor devices are installed at high density, the responses of spatially proximate sensors are highly correlated. In other words, these sensor responses indirectly include location information. Based on this consideration, we proposed a sensor localization method using the spatial correlation of nongeotagged sensor response data and evaluated its performance. Our method enables location estimation from the response data obtained during environmental monitoring, even for sensors that are not equipped with range functions. The simulation experiments showed that a more accurate sensor localization is possible with an error of 1 m using a wider sensor detection range, even with rough proximity pair ratio settings. In addition, we verified that there is a trade-off between localizability and observability during environmental monitoring, such as the visualization of vehicle positions from sensor responses. Our proposed method may be applied to future urban environments with a large number of sensor devices installed for environmental monitoring.