Low Complexity FFT Based Spectrum Sensing in Bluetooth System

In this paper, we propose a complexity reduced spectrum sensing technique for Bluetooth to find channels free from the use of other communication devices in 2.4 GHz ISM band. By exploiting the spectrum characteristics of interference sources, the proposed scheme detects the availability of chan- nels by comparing the power spectrum density (PSD) with a threshold. To reduce the implementation complexity, the PSD is calculated by means of fast Fourier transform and linear interpolation. The threshold for the detection is determined to maximize the detection probability. To further improve the trans- mission performance, the proposed scheme dynamically changes the channels by measuring the transmission performance. I. INTRODUCTION source nearby by measuring the PER. Adaptive control action mitigates interference sources based on the channel classifica- tion result. The AFH can outperform the RFH by only utilizing channels in good condition. However, it is required to blindly transmit packets for the channel classification, causing possible packet loss and PER performance degradation. Problems with the use of conventional RFH and AFH techniques can be alleviated by means of efficient spectrum sensing. The spectrum sensing is often achieved by three detection techniques: energy detection, matched filter coherent detection, and cyclo-stationary feature detection (7). Since non-coherent energy detection is simple and is able to quickly locate the spectrum occupancy information, it is widely used for the spectrum sensing. Previous works considered the en- ergy detection by cooperating among multiple radios (8), (9), but they can be applied to the detection of signals in a single channel. The signal detection in multiple channels can be achieved by estimating the power spectrum density (PSD) of a wideband signal. The PSD of a wideband signal can be estimated by means of fast Fourier transform (FFT). The existence of interference was detected by approximating the PSD of each channel as a Gaussian model (10). However, it may take a long time to get a desired PSD from the FFT results. When applied to Bluetooth, it may need to perform FFT for all Bluetooth channels, resulting in huge computational complexity. In this paper, we propose a complexity reduced FFT based spectrum sensing scheme to detect the existence of WLAN signal. The proposed scheme performs the FFT exactly once only for selected channels to cope with rapidly varying wire- less channel environment. The PSD of other channels can be estimated by a simple linear interpolation technique. Since the proposed scheme performs the FFT only once, the PSD of each channel cannot be approximated as a Gaussian model. In the propose scheme, it is described by using a probability density function (pdf) of each channel condition (i.e., busy or idle). The threshold for the spectrum sensing is determined by a maximum a posteriori probability (MAP) criterion to maximize the detection probability. The rest of this paper is organized as follows. Section II describes the system model in consideration. The proposed scheme for the mitigation of interference is described in Section III. Section IV verifies the performance of the proposed scheme by computer simulation. Finally, conclusions are given in Section V.