Research And Implementation Of Key Technology Of Host Computer For Adaptive Communication

With the vigorous development and popularization of various intelligent systems and smart devices,the data generated by real-time interacts exploded out of expectations,bringing more complicated interferences and challenges to current underperforming communication systems.An adaptive communication system which have the great performance of strong anti-interference,high spectrum utilization,and fast transmission provides a perfect solution to these chanllenges with adaptively changes communication conditions according to environmental spectrum usage,and utilizes spectrum idle gaps for communication.As a key part of the adaptive communication system,modulation mode detection is crucial for the final system to demodulate the correct information.Therefore,this paper applies machine learning to detect the modulation mode of digital modulated signals generated by different modulation modes of dynamic communication systems,including ten kinds of modulated signals such as MASK,MPSK,MFSK and MQAM,to help the subsequent information demodulation.The research and implementation of key technology of the adaptive communication upper computer is mainly to automatically identify the modulation mode of the received digital modulated signal with machine learning.Firstly,the modulated signals which generated by the cognitive radio platform with different modulation modes are obtained through the PCIE interface.Secondly,the characteristic parameters of signals are extracted with feature extraction methods.Then the Adaboost cascade classifier based on the Boosting thought is designed to classify the received modulated signals.In order to facilitate human-computer interaction and better reflect the beneficial effects of this algorithm,the corresponding PC software system is also designed for information management and results display.PCIE is used for the connection and data transmission between PC and FPGA board,while DMA is used for high-speed data reading and writing,and the speed can reach 3.44GB/s.Both system parameters and communication data are transmitted with PCIE interface and DMA protocol,making whole system achieve the exordinarily real-time performance.The main process of automatically detect modulation mode of digital signal includes feature extraction and signal classification.Since the amplitude-frequency-phase characteristics of the modulated signal express modulation mode more effectively,the high-order cumulant of the signal and the Haar wavelet transform are used to extract the amplitude-frequency-phase characteristics of the signal.Precisely because extracted signal features have intra-class linear approximation characteristics and more than two categories,this paper uses SVM as a weak classifier to construct an Adaboost cascade classifier for classifying the modulation mode of digital signal.And the simulation results show that the detection accuracy of this algorithm can reach above 95% when SNR>1d B,and it can reach 98% when the SNR>2d B,fully verifying the effectiveness of the proposed method in this paper.