GNSS Ranging Code Design Based On Genetic Algorithm

With satellite navigation experiencing an exponential increase in civil as well asmilitary applications, global navigation satellite system (GNSS) is playing a more andmore important role in various aspects of people’s lives. The key indicators of satellitenavigation systems, i.e. range resolution and ranging accuracy, mainly depend on theperformance of ranging codes. Therefore, design of the ranging code has been theresearch priorities of a satellite navigation system all the time. According to theirgeneration methods, ranging codes used by current GNSS can in general be divided intotwo categories, codes based on linear feedback shift register (LFSR) and codes based onoptimization algorithms. The former are prone to generate, but only exist in specificlengths and can’t meet the needs of future satellite navigation; the latter can be flexiblydesigned of any length, of which the most typical ones are Random codes that generatedbased on genetic algorithm (GA) used in Galileo system.Based on natural selection and genetic theory, GA simulates the reproduction,crossover and mutation phenomenon which occurs during the process of biologicalevolution, follows the principle of ‘survival of the fittest’, carries out iterative search andeventually obtains the global optimal solution. Due to its good search performance, GA iscommonly used for optimization problems. Design of the ranging code needs to seek thebest balance between its autocorrelation and cross-correlation properties, which is atypical optimization problem. And Random code has excellent performance, so this paperattempts to explore the design process of Random code using GA, in order to providesome reference information for the ranging code design of our COMPASS system.This paper firstly introduces the function of ranging codes, the generation method oftraditional ranging codes and the four ranging codes used in navigation systems, andpresents a brief analysis of their performance. On this basis, design requirements andevaluation criteria of ranging codes are proposed. Afterwards the mathematical basis ofGA is presented, including the schema theorem and the building block hypothesis, and itsbasic principles and workflow are studied in depth. Finally, the fitness function of GA isdesigned according to the evaluation criteria of ranging codes, and crossover andmutation operators are specifically designed based on its mathematical theory. To ensureexcellent performance of the obtained codes, a optimization scheme is implemented uponthe generated codes based on GA. Comparisons between the obtained codes and the GPSL1C/A code and Galileo E1OS Random code are carried out, and show good results,thus verifies the feasibility and effectiveness of the proposed algorithm.