Rtklib new version 2.4.3 B34 test comparison

2020/12/30 , At the end of last year ,rtklib Finally, a new version has been updated .clone Came down to test . Say first conclusion ,

1. The efficiency of the new software has been greatly improved 2 To 5 Between times
2. The accuracy can be considered unchanged
3. Removed some files , for example Some boards do not support
4. qt The project no longer supports , let me put it another way qt Project not deleted , But there will be no subsequent maintenance
5. app The catalogue has been restructured , More organized

The first set of comparative data

This set of test solutions is too weird , Deeply suspect that there is a problem , But put it here first

Test data use 148 Data collected continuously for hours , Collection conditions are poor , The solution effect is not very good . Here's the comparison .

In the process of collection , Split files every hour
The collected data are static observation data
Dynamic test can refer to a friend's comparison results Dynamic test comparison

rtklib edition	Average hourly solution time （ millisecond ）	Hours of successful solution
b34	263.75	102
b33	1816.77	18

Calculation time curve of two versions of software ,

Comparison of the whole week fixed success numbers of the two versions .

From the data collected over the past few days , The new version is much more efficient and effective .

The second set of comparative data

The first set of data is of poor quality , Not representative , So later I collected 40 Hours of good quality data for comparison test .

Comparison of solution time ,

The whole week ambiguity is compared with a fixed number of times per hour ,

Let's take a look at the mean square error of the two versions of calculation data

version	std_x(m)	std_y(m)	std_z(m)
b33	0.015925791	0.034312591	0.032847452
b34	0.015927796	0.034271672	0.032787868

From this set of data , Whether it is the whole week fixed number or the mean square error of the final solution , The precision difference between the two versions is very small .