Code Sequence Generation with Genetic Algorithms, with Correlation Properties Similar to GPS C/A Codes

Abstract

C/A codes (course acquisition codes) are pseudo-random generated codes with good correlation properties. Those codes are being used in GPS. Each satellite vehicle can generate its unique C/A code sequence and modulate it with the output data signal. Each millisecond of satellite data contains 1024 chips (bits) of C/A codes and each 1ms this code sequences are being repeated. Receivers are using locally generated C/A codes in to filter out the satellite signal from aggregated signals near the receiver’s antenna. As the C/A codes are being repeated in each 1ms, in theory, it’s enough of 1ms satellite signal to determine from which satellite it is coming. C/A code’s correlation properties are being used for filtering incoming signals. The higher is the autocorrelation properties of the code sequence, the easier to filter it out from the summary signal. In other words, the same C/A codes have high correlation values and different C/A codes have low correlation values. The target of this article is to write a program using genetic algorithms that will generate code sequences from 1 and −1 values that will have nearly the same correlation properties as the C/A code, where each individual in algorithm will contain 32 number of code sequences with 1024 length that has low cross-correlation and high autocorrelation properties.