Yesterday we sat down and cut-and-pasted together a simple Linux C interface to the OpenShoe. This allowed us to logg data at full speed (820[Hz]) from two units simultaneously. This we have had problems with doing reliably from Matlab. The problem seem to be that Matlab have a very large overhead in its fread-calls. This can probably be solved somehow by using large buffers, reading a lot of data at the same time, and do some clever parsing. This is how we solved it for a single units. However, for two units we found it easier to just write a C interface. This also allowed us to reuse some old synchronization code of ours. Currently, the C interface is kind of uggly, untested, undocumented so we have not made it openly available yet. Possibly we will do this in the future. If you have an urgent need for such an interface and do not wish to code one yourself (if you have a look at the corresponding Matlab control scripts, it is rather straight forward), contact us and we will see what we can do.
Archive for January, 2012
It was recently discovered that the matlab-code for the Algorithm test framework was missing from the sourceforge repo. This has now been added. Apart from communication scripts, this code also contains an intermediat version of the algorithms. It is in matlab code like the OpenShoe matlab implementation but it is written only component wise multiplications like the C version in the Navigation algorithms repo.