Data Driven Environmental Awareness Using Wireless Signals

TL;DR

The paper tackles robust indoor/outdoor classification of RF environments to enable spectrum sharing in the 6 GHz band. It trains and evaluates a deep neural network on a comprehensive SigCap dataset that fuses Wi‑Fi, cellular, and GPS features to classify three environments: Outdoor, Indoor Interior, and Indoor Near Window. The authors show that extended windowing and majority voting substantially improve accuracy, with DNNs delivering the best performance and near-perfect results when training data are diverse. The work provides a public dataset and outlines practical steps toward on-device implementation and time-series models, highlighting significant potential for improving spectrum coexistence and interference management in shared bands.

Abstract

Robust classification of the operational environment of wireless devices is becoming increasingly important for wireless network optimization, particularly in a shared spectrum environment. Distinguishing between indoor and outdoor devices can enhance reliability and improve coexistence with existing, outdoor, incumbents. For instance, the unlicensed but shared 6 GHz band (5.925 - 7.125 GHz) enables sharing by imposing lower transmit power for indoor unlicensed devices and a spectrum coordination requirement for outdoor devices. Further, indoor devices are prohibited from using battery power, external antennas, and weatherization to prevent outdoor operations. As these rules may be circumvented, we propose a robust indoor/outdoor classification method by leveraging the fact that the radio-frequency environment faced by a device are quite different indoors and outdoors. We first collect signal strength data from all cellular and Wi-Fi bands that can be received by a smartphone in various environments (indoor interior, indoor near windows, and outdoors), along with GPS accuracy, and then evaluate three machine learning (ML) methods: deep neural network (DNN), decision tree, and random forest to perform classification into these three categories. Our results indicate that the DNN model performs the best, particularly in minimizing the most important classification error, that of classifying outdoor devices as indoor interior devices.

Figures (9)

• Figure 1: Example locations representing outdoor, indoor interior, and indoor near window environments.
• Figure 2: Map of data locations.
• Figure 3: CDF plots of input features.
• Figure 4: Indoor accuracy by model and feature set, two-classes classification.
• Figure 5: Outdoor accuracy by model and feature set, two-classes classification.