readme

此程序为GPS接收机的源码

README the readme file for gpl-gps, a port of OpenSource GPS for the GP4020
 Copyright (C) 2005  Andrew Greenberg
 Distributed under the GNU GENERAL PUBLIC LICENSE (GPL) Version 2 (June 1991).
 See the "COPYING" file distributed with this software for more information.

GPL-GPS
Open source software for Commercial, Off-the-Self Global Positioning System 
(GPS) receivers. <http://gps.psas.pdx.edu>

PURPOSE
To make high quality GPS software available to anyone for experimentation,
education, research, and new applications. Our software is intended to run on
inexpensive off-the-shelf commercial hardware while attaining the full accuracy
possible with the present civilian GPS configuration.

Possible applications include but are not limited to

  Differential GPS featuring centimeter accuracy
  Improved accuracy WAAS GPS
  GPS aiding and complementary inertial navigation integration
  Carrier phase for vehicle attitude determination
  Multipath rejecting and jamm-resistant algorithms
  Software GPS
  GPS integration, for example as a subroutine in a software radio
  Development of new GPS systems such as L5 capable receivers
  A tool for teaching about the GPS

STATUS (alpha)
June 2005, gpl-gps runs from SRAM on re-flashed MG5001 OEM boards and
produces position fixes about 2 times less accurate than state-of-the-art
commercial software. This is thought to be due to 3 shortcomings with the
current implementation of gpl-gps. The most obvious missing piece is carrier
phase aiding, which provides an independent and ideally very accurate estimate
of receiver velocity. Also missing, a Kalman filter, and some re-working of the
fine tracking algorithm and it's interaction with local time estimates.

If these 3 shortcomings were addressed, we believe that our accuracy would be comparable to current commercial implementations.

(Some of the following is redundant with the TODO file.)

In addition to position computation issues there are several other features
that need work including: almanac/Warm start code, re-acquire tracking,
the search and acquisition algorithm, cold-start sky search code, and UI.

Features aside, the code base could benefit from some work. Modularizing,
possibly even dynamic loading, porting to other chipsets, software correlators,
a command interpreter. Support for fixed point, efficiency improvements, all
the usual suspects.

INSTALLATION
Could somebody write this up? Maybe a separate file called "INSTALL"

HISTORY
Around 1995 Clifford Kelly began working on open source GPS software. He called
the project OpenSource GPS (OSGPS) and as of 2005 it has reached version 1.17.
The project website is <http://home.earhlink.net/~cwkelley>, there is also a
project page on sourceforge <http://sourceforge.net/projects/osgps>.

By using a hacked commercial receiver based on the GP2021 chip, bypassing the
receiver's microprocessor and connecting the hardware correlator directly to
the PC-ISA bus, OSGPS is able to run all its software on the host PC.

Meanwhile, at Portland State University, a small group of students started a
rocket club, the Portland State Aerospace Society (PSAS) whose admittedly
ambitious purpose was to (eventually) send nano-satellites into orbit. Through
the late 1990's this group <http://psas.pdx.edu> was semi-seriously discussing
the idea of open source software for inertial navigation with GPS aiding based
roughly on the complementary filter approach.

Finally in the year 2000 Andrew Greenberg, a PSAS member, started what became
the gpl-gps project. As circumstance unfolded, not much actual software got
written at first. The project moved slowly for several years until April 2003
when the gps.psas.pdx website was finally set up. This was followed by a flurry
of activity culminating in a meeting at the 2003 Institute of Navigation
conference. After the conference many distractions sapped the energy away from
GPS work and the project languished without progress.

Then in 2004 Takuji Ebinuma ported OSGPS to the GP4020
<http://www.geocities.com/tebinuma/osgps>. Unlike Cliff Kelly's effort,
Takuji's port ran on unmodified OEM hardware using the GP4020's built in
bootloader. Though both approaches have advantages, the off-the-shelf hardware
approach of Takuji's effort was more in keeping with the goals of gpl-gps, and
more to the point, Takuji was making progress. This, combined with the 
realization that his graduate credit hours would soon begin to expire,
Andrew leaped into action and in April 2005 gpl-gps was revivified.

By early May 2005 gpl-gps had been successfully ported to the SigTec MG5001 OEM
board using Linux as the development platform. By the end of May, his thesis
finally complete, Andrew was awarded his long sought masters degree.

Currently, gpl-gps is being reworked. We hope to rapidly improve its
performance so that it is comparable to similar commercial products. This will
involve adding several features, improving certain existing components, and
unifying and streamlining the code base.