Studies And Applications Of Some Key Techniques In Radio Monitoring Based On Particle Filtering

Radio technique is one of the most developed techniques in the world, which plays a more important role in promoting the progress of social economy and improving the standard of living of the people. With the development of variant radio techniques and the abundance of service types, the electromagnetic environment in the air becomes more and more complexity. The transmitted radio waveforms that have the characteristic of openness could be disturbed by other electromagnetic waveforms or malicious attacks. In this situation, the advanced and reliable radio monitoring techniques are required to be as the security guarantee. In radio monitoring work, the key techniques of interest are parameters measurement of unknown signals, elimination of malicious disturbed signals and localization of interferers. This dissertation is on the basis of signal processing theory. Some modern signal processing methods, especially the particle filtering-based techniques are proposed to solve the problems of signal parameters estimation, channel equalization and interference cancellation, and signal source localization. The main researches and conclusions are listed as follows:(1) The estimation problems of the signal parameters, such as carrier frequency, baud rate, time delay and frequency offset, are studied. The performance of the baud rate estimation method proposed by Chan is limited by the carrier initial phase. We propose the joint baud rate and carrier frequency estimation algorithm based on phase-locked loop (PLL) and wavelet modulus maxima. The method uses the PLL to synchronize the carrier frequency and abstract the signal envelope. The frequency output is processed by Kalman filter to obtain the optimal estimation of the carrier frequency. The phase output is processed by wavelet transform and the modulus maxima are kept for baud rate estimation. The traditional particle filtering-based time delay estimation algorithm is inefficient for small delays. We propose an estimation algorithm based on forward and backward detection for time delay and frequency offset joint estimation. The particles are adaptively sampled to adjust the variance of the state noise to provide high estimation accuracy.(2) The problems of blind equalization and single antenna interference cancellation are studied. The radio signals transmitted in space could be disturbed by inter-symbol interference (ISI) and the co-channel interference (CCI). The performance of signal identification and detection would be severely degraded. According to the above cases, we propose the particle filtering-based time-varying multi-path blind channel equalization and single antenna interference cancellation algorithms, respectively. We consider Alpha-stable distribution noise as the additive noise, and use Gaussian distribution to approximate it. The method uses the state-space to model the channel coefficients, noise parameters and transmitted symbols. The joint posterior distribution of these variables is recursively calculated to track the fast fading channel and detect multiple signals in the presence of single antenna.(3) The problems of signal source localization are studied. We focus on three application problems of radio monitoring, which are bearing-only signal source location based on mobile station, interferer localization based on scattered signals from airplanes, and sensor nodes localization in wireless sensor networks. According to particle filtering theory, we propose a new nonlinear filtering algorithm mixed by particle filtering and Kalman filtering for parameters estimation and bearing-only location, propose Gaussian approximation particle filtering algorithms in the presence of Alpha-stable distribution noise for interferer localization by using the Doppler frequency shifts abstracted from the scattered signals, and propose the wireless sensor nodes cooperative localization based on recieved-signal-strength measurements.