The Optimization of Algorithm and Implementation for Dual-Constellation Navigation Receiver

Tianyi Ma,Xiaowei Cui,Mingquan Lu,Zhenming Feng
2010-01-01
Abstract:Current dual-constellation GNSS navigation receiver can not fully satisfies the development of navigation industry. The architecture of current Compass and GPS dual-constellation navigation receiver is usually built upon FPGA and float-point DSP. Compared with fix-point DSP and FPGA, float-point DSP works slow and costs much, and it is also difficult to integrate float-point DSP into ASIC or ASSP. So in this paper, we discuss the optimized algorithm and the new implementation that integrate all parts of navigation receiver on one FPGA or ASIC.Our revised algorithm of navigation receiver is introduced first, mainly including acquisition and tracking mechanism. For acquisition module, we divide the whole acquisition process into three units: noise threshold detection, pre-acquisition, acquisition verification. For tracking module, with the help of accurate fix-point simulation, float-point variables are converted to fix-point ones, and rescaling is applied during all process to assure precision.Algorithm only makes sense when it combines with implementation, and the quality of algorithm also should be tested by application. So we also introduce the implementation of navigation receiver baseband algorithm on FPGA platform, including platform architecture, module division, work flow, and some optimization mechanics including resource sharing, process combination and etc. And we simply introduce the measure result and the application of dual-constellation navigation receiver baseband processor.And viterbi decoder for navigation receiver is also discussed. This paper brings forward a revised mechanism of viterbi decoder especially for navigation receiver. We borrow the soft output viterbi algorithm, but do not adopt the algebra ring method, and put forward revised viterbi mechanics. And the optimization is also on shorten-decoding, survival-path combination, butterfly calculation of ACS unit.
What problem does this paper attempt to address?