Abstract: In the cognitive Internet of Things (CIoT), due to the non-cooperative characteristics between the primary user (PU) and the secondary user (SU), it is unreliable to seek the spectrum access opportunity by merely relying on traditional spectrum sensing technology. As an important auxiliary information, the mutual location information between PU and SU can assist in determining the possibility of secondary access to the licensed frequency band (LFB). This paper proposes a low-complexity neighborhood-based weighted centroid localization (NB-WCL) algorithm based on received signal strength (RSS) measurement and the strategy of conforming neighbor relationship. By solving the SU’s localization problem in the CIoT, we can accurately set the LFB-access enabling flag for each SU in the CIoT. This paper also analyzes the root mean square root error (RMSE) performance of two-dimensional position estimation and verifies the impacts of factors such as communication radius, node density, shadowing influence, path loss exponent, connectivity metric, and the number of data transmitted on the algorithm performance in simulations. The theoretical derivation and experimental results show that the proposed scheme provides more robust and better localization error performance for the SU localization algorithm in CIoT than the traditional localization algorithm, which can effectively enhance the reliability of cognitive IoT for spectrum access. The proposed scheme can serve as a practically effective candidate solution in the CIoT.

Key words: weighted centroid localization, cognitive Internet of Things, neighbor relationship, performance analysis, spectrum access