Distributed sensing of spectrum occupancy and interference in outdoor 2.4 GHz Wi-Fi networks

A spectrum monitoring campaign was launched in an outdoor urban radio environment to investigate the potential deployment of Cognitive Radio (CR) Wi-Fi networks in the 2.4 GHz ISM band. The campaign used a CR learning platform with 4, 8, and 16 sensors. This paper presents Wi-Fi spectrum occupancy and interference behaviour based on the outcome of one of the measurements using 16 GPS-synchronized sensors. At detection thresholds T_d ≥ -62dBm, long-term spectrum holes were dominant on all Wi-Fi channels. However, at T_d <; -62dBm channel availability decreased making CR operation more challenging. Wi-Fi traffic was dominated by management packets which correlated strongly with channel occupancy and far exceeded data and control packets. We noted management and data packet redundancies in the current IEEE 802.11 standard causing inefficient spectrum utilization. About 2/3^rd of management packets were beacons, and more than 1/2 of data packets were data-reserved and null (no data). We also noted, at T_d ≤ -82dBm, a large number of Wi-Fi users more than can be attributed to the immediate surroundings of sensors, but a small set of them was dominant producing the bulk of spectrum occupancy and interference. In addition, at least 25% of these users were detected only once over the 5.5 hours measurement time span. The measurements showed spatial variations of the Wi-Fi environment over a small sensing area (21m-by-45m). We noted considerable non-homogeneity in the distribution of Wi-Fi interference at T_d ≥ -62dBm, but some non-homogeneity at T_d ≤ -82dBm. We also noted significant correlated fluctuations of received signal strength and non-reciprocal links over short distances between sensors.