multichannel discrete observations

[Dina Daitchman]

Hi,

I would like to perform HMM with multichannel categorical data,

I mean that  every minute I have data from multiple channels, for example speed, acceleration and step counter.

each channel is binned, so the possible values are from several discrete distributions 

How can I model this case  with the JAHMM package?

Thank you,

Dina