Abstract de la publi numéro 10425

The current GPS provides an open positioning service by broadcasting the L1 C/A signal. This signal allows precise positioning in environments with a high C/N0, i.e. where the direct GPS signal is received without any significant attenuation, with no obstacle on the line of sight between the GPS satellite and the user receiver, and without multipath. However, without assistance, in indoor and in narrow urban environments the previous statements are no longer valid and the C/N0 is not always high enough to allow a correct data message demodulation/decoding or a satisfactory tracking process. The introduction of two new signals, GPS L2C and GPS L5, has improved this capacity of retrieving the correct message; the information without errors. These new signals are constructed with a serial implementation of channel codes: first, a convolutional code (171,133) used to implement the FEC (Forward Error Correction) and, second, a new and more performing code, the CRC-24Q, aimed at detecting words with uncorrected errors. In a serial implementation we call the first code the inner code, in this case the convolutional code, and the second code the outer code, in this case the CRC-24Q. However, although these signals fairly improve the decoding performance compared to GPS L1 C/A signal, they still do not provide an ideal service: for any C/N0, the BER can be improved. Therefore, the aim of the method presented in this paper is to reduce the C/N0 required by a GPS L2C/L5 receiver to obtain a determined BER value during the decoding process compared to the C/N0 levels of the traditional decoding techniques. This method reuses the imposed data message structure in a different form exploiting one of the only available degrees of freedom provided by this defined system: the data processing. More specifically, the proposed method combines the two message channel codes (inner and outer) in order to develop a new decoding process performing better than the conceived serial implementation of the two individual decoding processes. The main idea of the method is to search through a modified Viterbi Algorithm for the most probable sequence being emitted that verifies the CRC-24Q. Indeed, this latter code determines if the sequence belongs to the coded words alphabet.