Channel Estimation In Wireless Local Area Networks And Cognitive Radio Technology Research

With the increasing scarcity of radio resource, Cognitive Radio (CR) enjoys tremendous attention for the reason that CR is a potential solution for contradiction between spectrum allocation and usage. Currently, CR has been applied in IEEE 802.22 Wireless Regional Area Network (WRAN), which utilizes CR to detect spectrum hole in TV broadcast band and then uses these spectrum holes to communicate without interfering TV and Wireless Microphone, thereby increasing spectrum efficiency.Since channel estimation, one key technique in CR, plays a vital part in CR realization, this paper focuses on two directions in channel estimation: (1) in order to obtain the excellent performance of Orthogonal Frequency Division Multiplex based Transform Domain Communication System (OFDM-TDCS) in low signal to noise radio (SNR) when it is applied to WRAN, we research how to eliminate impact of noise especially when SNR is very slow and improve the performance of channel estimation in WRAN; (2) in order to improve the performance of channel estimation as well as save radio resource and transmitting power in CR system, we strive to find much effective Bayesian semi-blind channel estimation method.According to the two research directions, the main contents of this dissertation are also split into two parts:Several channel estimation methods based on noise reduction are proposed for OFDM-TDCS, which performs well when SNR is very low. In time domain, time moving average, time forgetting average and the combination of time average and time forgetting are proposed to eliminate the impact of noise by using the slow variation characteristic of WRAN slow fading channel. In IDFT transform domain, low-pass filter based method is applied. Simulation results show that the method is effective in WRAN slow fading channel; the performance of OFDM-TDCS using practical channel estimation is only a little inferior to that using ideal channel estimation.Corresponding to the second research direction, we first summarize the current Bayesian methods applying to semi-blind channel estimation, including Kalman Filter (KF), Particle Filter (KF), and Mixture Kalman Filter (MKF), and then summarize the current channel prediction models. Based on the current channel prediction models, we proposed a new dynamic channel prediction model and adaptive estimated the parameters of the model by applying the sequential evidence maximization with sequentially updated prior method.Finally, a new joint data detection and channel estimation scheme is proposed for Alamouti space-time block coding (STBC) systems. In this scheme, we utilized the new dynamic channel prediction model we proposed and the corresponding channel estimator based on adaptive Kalman filter. Therefore, the estimation of maximum Doppler frequency shift is not required in our proposal. Compared with the traditional Kalman estimator, the proposed estimator has better performance and lower complexity. Simulation results show that the proposed estimator has stable performance under different maximum Doppler frequency shift.