Previous Abstract Return to Session A1 Next Abstract

Session A1: Algorithms for GNSS Processing and Sensor Integration

Acquisition of 3 GNSS Signals of GPSL1CA, GPSL1C and GalileoE1OS Simultaneously in a Single Processing Chain that Halves Processing and Battery Power
Ali Albu-Rghaif, College of Engineering, University of Diyala, Iraq; Ihsan Alshahib Lami, The University of Buckingham, UK
Location: Grand Ballroom G
Date/Time: Tuesday, Jan. 30, 3:05 p.m.

Peer Reviewed

The GPS L1C signal is designed to improve on the localization accuracy achieved by the existing GPS L1CA and Galileo E1OS signals. All these three signals are interoperable with each other and we believe that designing an integrated solution to acquire these three signals in a single processing chain will save much valuable processing time and energy, especially when this solution is deployed in a mobile or a battery-less device. Our novel solution “CACOS” proposed here uses the Compressive Sensing technique to acquire all three signals simultaneously in a single CS-framework. Our simulation results show that CACOS achieves (1) better acquisition performance, (2) halves the implementation/resources size, and (3) halves the energy consumption required for simultaneously acquiring these signals, when compared to most GNSS receivers used onboard Smartphones.
CACOS implementation consists of three stages:
1. The three signals are sampled based on the Bandpass Sampling receiver technique with a single folding frequency.
2. Multiply the sampled signals (n length) by a range of Doppler channels (m length) simultaneously to generate nonDoppler shifts vectors, where the channels with the best match to the frequency of the received signals will then be selected for our CS-framework.
3. In this CS-framework stage, CACOS compress the information of all the signals to a matrix, where the compressed information is then matched with the combined dictionaries/correlators of all three GNSS signals to determine the code-phase delay of the acquired SV as well as the Doppler frequency shift.
Note that, all the required matrices such as the Doppler channels matrix, sensing matrix and the dictionary matrix are generated once only and are then stored in local memory for use during the acquisition process. MATLAB simulation results show that the “computational complexity” of our CACOS implementation is one-third of that of traditional implementations such as when using Matched Filters. This reduction is achieved because the input to CACOS CSframework is transferred to a “compressed format” and therefore the acquisition process is now dependent on the number of Doppler channels rather than on the dwell time. In addition, CACOS performance is as good as that of a Matched Filters implementation, but can achieve higher frequency resolution acquisition to about 20Hz that equals to doing fine frequency search.



Previous Abstract Return to Session A1 Next Abstract