Frequency of Gyroscope
I am using 400 Hz to collect data on the MMR sensor, however when I export the data from the MetaBase App to my computer and read it into MATLAB, the time values are not consistently spaced. There are 405 data points per second and the difference between subsequent values ranges from 0.003s and 0.001s. Over a large range this might not be a huge concern, but when integrating the gyroscopes angular rate data over a very small time step, this discrepancy could make a big difference in the angle calculated. Does anyone have any comments or advice regarding this issue?
Comments
Are you streaming or logging?
I am logging, not streaming.
Does this make a difference?
I don't see any issues when I log at 400Hz on my MMR.
Can you elaborate? Post data, be more specific about expectations versus reality and so on.
Here is a screenshot of the data exported from the MMR into Excel. I expected the time difference between data points to be 1/400s (0.0025s), however this is not the case. Under elapsed (s) column, some data points appear to be collected 0.001s apart, others 0.002s apart, and others 0.003s apart thus the sensor is not sampling at consistent frequency.
Take a look at the epoch instead, don't use the time elapsed.
Also note this is all within the acceptable error rate for bosch sensors + bluetooth latency.
Is there any reported issue with the Epoc data when using the recent firmware versions on MMS devices?
Epoc incremented normally with time when I look at my data collected in June 2021 (log mode). I do not remember the firmware version at that time.
However, the Epoc do NOT increment when looking at my data obtained from November 2021 (the time stamp DO). Log mode as well. I do not have the sensor with me as I am writing this message, but I have regularly updated the firmware version and I remember I did it before starting a series of experiment in November 2021 (1.7.2 or 1.7.3 if I remember well)?
I had written a small script to rephased and merge acc and gyro sensor data based on epoch in June 2021. Of course I can do te same using timestamp info, but I did not notice the issue until recently and found that all my data collected and treated after Nov. 2021 may be corrupted:(.
Can you confirm that the Epoch data has been corrupted since the recent firmware update? or maybe it is my mistake?
Sorry for the message. I just found some of my dataset have a normal Epoc. The issue does not seem related to the firmware version. I will dig into it to understand the issue and post an update if I can find. m(_ _)m.