The Research On Multiuser Detection Based On Hopfield Neural Network

Code division multiple access (CDMA) communication system possesses many attractive properties, such as powerful capability of interference-resistance and large system capacity. But the existence of multiple access interference (MAI) limits the capacity and performance of CDMA system. Therefore, Multiuser detection is proposed to suppress the MAI. By far, most multiuser detection algorithms have flaws such as algorithm sophisticatedly, low speed of convergence and so on, but neural network have many merits such as high computing speed, strong parallel processing capacity, thus it becomes a research focus in the study of multiuser detection. This paper analyzes optimum multiuser detection, linear detection, decorrelating detection and some other detection strategies, and investigates multiuser detection based on Hopfield neural network.The main work of this thesis can be divided into three parts: The first part of this paper introduces the knowledge of mobile communication system and multiuser detection interrelated; the second part is mainly about the basic theory of Hopfield neural network; the last part of this paper is the investigation, simulation and analysis of multiuser detection based on Hopfield neural network in mobile communication system. Thereinto, this paper places its emphasis on the last part; its main contributions are as follows:1. Make analysis on the optimum detection based on HNN, as to the drawback of the detector, this paper proposes a modified scheme which has better Bit Error Ratio (BER) performance and faster convergence rate than original detector verified through the simulation.2. Combine linear multiuser algorithm based on MMSE criteria with Hopfield Neural network, this paper presents a linear HNN detector based on MMSE criteria, reduces the computational complexity of linear MMSE multiuser algorithm. The simulation results show that it gains better BER performance.3. This thesis investigates the optimal multiuser detector based on HNN in MC-CDMA system. The simulation results show that in MPF chan -nel by the use of perfect CSI, the optimal multiuser detector based on HNN can gain suboptimal performance in MC-CDMA system, and the availability of the CSI to the detector is very important. 4. This paper applies the blind multiuser detector based on HNN into MC-CDMA system, and proposes KCNN blind multiuser detector and modified blind multiuser detector based on HNN for MC-CDMA system respectively. The simulation results show that both detector can suppress the effect of MPF and possess better BER performance and channel -tracking performance. Moreover, the modified blind multiuser detector based on HNN is superior to KCNN detector.