Study On Key Technology For GNSS Antenna Array Signal Processing

As the applications of GNSS continue to expand, the vulnerability of its anti-jamming performance has become an important issue. The antenna array processing is a promising technique to overcome the vulnerability problem of the GNSS receivers. In recent years, the array processing theory has been widely studied for GNSS receivers. However, most of these researches are the extension of the processing techniques developed for radar and communication areas and have no optimization for GNSS characteristics. This paper focuses on the anti-jamming and low cost requirments of GNSS receivers, and studies the joint space-time processing and blind beam-forming techniques as follows:(1) The optimum criterion for space-time processing for anti-jamming and beam-forming is to maximize the equivalent C/No, while this criterion can not be implemented in real application as it is difficult to obtain the prior information needed by this criterion. To replace the optimum criterion with a realizable criterion with high performance approximation, a multi-object programming algorithm is adopted to consider two optimization goals which are depression of jam and loss of signals. The algorithm formulate the two goals in convex function form, by introducing an object weight coefficient the optimization effort of the two goals can be adjusted, and the method for finding the optimal object weight coefficient is then developed. Theoretical analysis and simulation shows that the algorithm can approximate the optimum maximization C/No criterion and can deal with complicated scenarios that traditional algorithms failed.(2) Inversion of the covariance matrix is a b asic operation in realization of space-time anti jamming algorithm, and its computation load grows rapidly with the number of time taps, therefore its application in the mass GNSS receiver is refined. Aim at this problem, a block toeplitz inversion algorithm is adopted into the anti-jamming processing. By improving the estimation method of the covariance matrix, the algorithm gets a sample covariance matrix both in the form of block Toeplitz matrix and Hermite matrix, and then the fast block Toeplitz inversion algorithm is applied and reduces the the inversion complexity of covariance matrix of K taps from O[K3] to O[K2]. With 4 antenna elements and 9 tap, the anti-jamming performance difference is less than 1d B, and the computation load reduction is 2/3 times of the traditional algorithm.(3) The blind beam-forming algorithm has limited DOA estimation accucary. A novel joint array tracking algorithm is developed to estimate the DOA information which is vital for beam-forming algorithm. The algorithm is based on a novel decomposition model of the dynamics. In this way, the joint vector tracking theory that is developed for single antenna system is extended to antenna array system.The common dynamic stress caused by transportation has the same impact on every elememt and a virtual single elem ent is proposed to adopt the vector tracking theory. T he dynamic stress differences between elem ents are caused by rotati on around the virtial element. The sm all dynamic stresses introduced by ro tation are tracked by accurate narrow-band loops. Meanwhile, the carrier ph ase differences derived from those loops contain all the information of DOA estimation, which is useful for beam for ming. The algorithm first uses the joint vector tracki ng to adopt high dynam ic stress with high sensitivity, and then the DOA related carrier phase information is obtained by accurate tracking loops. Compared with separate tracking for each individual loops, the DOA accucary can be improved by 3 times.(4) The data modulation prohibits long time coherent accumulation that is essential for improving tracking sensitivity. Conven tional method searchs all data m essage combination and the computation increases exponentionally. An approximate algorithm with controlled com plexity is developed and used in high perfor mance FLL discriminator. By using the dynamic programming philosophy, this algorithm can predict and exclude navigation data bit combinations in the c orrelation value accumulation process with quantization the angle of correlation vector, thus reduces the algorithm’s complexity to be in lin ear relationship with data bit leng ths.The linear complexity makes it possible to implem ent a 0.12 s coherent integration for the 500 bps signals, which increases the FLL sensitivity for 7d B.(5) The joint carrier phase tracking architecture for an tenna array processing has high sensitivity for the common dynam ic tracking as the vector tracking is adopted. However, although the rotation dynamic is relatively small and introduces little dynamic differences between elements, it is dif ficult to work at the sam e low C/No as the joint vector tracking loop. To solve this problem, the bit estimation algorithm for low C/No is used in the differential dynam ic tracking. Thus the conflict between dynam ics and sensitivity of tradition al loops is solved. Particularly, a combin ed tracking algorithm that uses FLL to track the common dynam ics and uses high sensitivity PLL to track the differential dynamics is proposed. The high sensitivity PLL adopts the f ast bit message estimation algorithm The com bined structure take into account the possible high dynamic and the DOA related carrier phase m easurement. Simulation results show that for a 4 element processing, stable DOA estimation can be obtained as low as 25 d BHz.The research and application s are summari zed at last. So me of the results have been used in various airborne and missile-borne anti-jam receivers.