Revised PRN code structures for GALILEO E1 OS

Pseudo Random Noise (PRN) Codes are a fundamental element in any Code Division Multiple Access (CDMA) system such as GPS, Galileo, QZSS and Compass. In fact, these codes are the tool that enables a GNSS receiver to distinguish one satellite from another. The current codes of Galileo for E5a/b, E6 and E1 OS were defined and selected in 2004. At that time, different approaches for code generation and optimization were considered. After the evaluation of all code sets, the Random Codes (available in the SIS-ICD [2]) were shown to provide the best performance for Galileo E1 OS according to the set of criteria defined at that time [6]. As shown in those papers published by the people involved in the Galileo code design at that time, the code families of each particular band were optimized taking into account only code properties. This means that the real modulation characteristics of the signal, that is their particular spreading waveform and multiplex, were not considered in the code design. In fact, the derived codes of E1 OS would only be optimal in the wide sense if the data and pilot signals were transmitted in quadrature. However, as we know from the Galileo SIS ICD [2], the data and pilot components constituting the Galileo E1 OS signal will be transmitted in phase using a modified Interplex modulation. In other words, a typical receiver will see either the sum or the difference of data and pilot code. In this paper a new set of random codes is presented. These are optimized for Interplex modulation and offer thus an interesting improvement in performance regarding the main figures of merit for PRN codes, namely the auto-and crosscorrelation. Compared to other PRN code families which are derived following a certain generation algorithm and can be generated in the receiver at run time, Random Codes like those of Galileo E1 OS need to be stored in memory at receiver level as they are not based on any generation scheme in general. This introduces additional complexity and costs especially for low-end mass-market receivers. This paper will introduce an interesting alternative family based on Sidelnikov/Lempel-Cohn-Eastman (SLCE) sequences that fit exactly the desired code length of 4092 chips. These sequences do not require any truncatation or addition of any chip. Furthermore, this new code family offers extremely good autocorrelation properties. Especially for low cost receivers, a cheap and thus memory saving code generation unit as well as s rapid acquisition are of great importance and very desirable. An application of this concept to the Galileo E1 OS codes would however not be optimum since shortening the 4092 chips long Galileo E1 OS codes would imply a considerable loss of sensitivity as non-desired correlation side peaks would appear in the partial correlations. An alternative solution could be to introduce shorter code structures within the long code of 4092 chips, such that the receiver would only concentrate on the long or short parts of it depending on its particular requirements. In order to introduce this substructure into the 4092 chip code, with which the correlation for the low end receiver could be accomplished, the concept of majority voting will be applied under the constrain of guaranteeing the correlation properties of the long 4092 chip codes.