Wireless Access

Reply
Contributor I
Posts: 62
Registered: ‎01-07-2015

How can I train VisualRF?

[ Edited ]

Decided to re-write this post as no responses...

 

I want to train visualRF as I'm not happy with the accuracy of the heatmaps we're seeing. However this doesn't seem possible in the HTML5 ui. 

 

If I switch to the flash ui all my floorplans break (truncated image, nothing lines up any more). 

 

So is there a way to train VisualRF from the CLI, or some other way?

Contributor I
Posts: 62
Registered: ‎01-07-2015

Re: How can I train VisualRF?

Well I'm guessing this isn't possible then. Shame as our visualrf heatmaps are somewhat optimistic.
Moderator
Posts: 1,252
Registered: ‎10-16-2008

Re: How can I train VisualRF?

In the Flash version of VRF, there is a survey feature that allows you to take surveys from static clients that would then assist in providing floor plan attenuation data for location calculation.  This feature has not yet made it into HTML5, and is on the list of features still to be ported over.

 

If you don't want to swap back to Flash, you can make use of the AMC client (check it out on your AMP's Home -> Documentation page).  The AMC client runs in the background of Windows laptops and uses the laptop's wifi to collect data and report back to AMP (similar behavior to APs in sensor mode).

 

The swap between Flash and HTML5 should not affect AP location.  That sounds like a defect.  Please open a support case to have support take a look.


Rob Gin
Senior QA Engineer - Network Services
Aruba Networks, a Hewlett Packard Enterprise Company
Contributor I
Posts: 62
Registered: ‎01-07-2015

Re: How can I train VisualRF?

Thanks Rob. I'll have a play with AMC. I've also opened a case about the problem switching between ui versions.

Contributor I
Posts: 62
Registered: ‎01-07-2015

Re: How can I train VisualRF?

So it would appear that creating floor plans on html5 and then switching to the flash ui doesn't work. Understandably the official response is that no development time will be spent on the flash ui. That's fair enough. Hopefully the missing functionality will appear on html5 soon.

 

I'm not clear on how to use the AMC to improve VisualRF accuracy. There needs to be a way of telling Visual RF where a client is, rather than it placing the client on the floorplan where it reckons it probably is... which is invariably wrong in our case.

 

Any insight most welcome.

Moderator
Posts: 1,252
Registered: ‎10-16-2008

Re: How can I train VisualRF?

The AMC client combined with the survey option from the Flash version of VRF can add to accuracy.  For the HTML5 side, the AMC client just provides added signal data for calculations.  Typically, you choose something stationary that has wireless capability, and set it as a static client.  If it's a mobile device, then you'll want to make sure to unset the static training if the device moves.


Rob Gin
Senior QA Engineer - Network Services
Aruba Networks, a Hewlett Packard Enterprise Company
Contributor I
Posts: 62
Registered: ‎01-07-2015

Re: How can I train VisualRF?

Thanks Rob, that's clearer... Essentially it isn't possible to do what I want until the html5 ui is feature complete. I can only hope that happens soon.
Dax
Occasional Contributor I
Posts: 6
Registered: ‎11-28-2012

Re: How can I train VisualRF?

A year and a half later and still no static training in html5.  Is it unneccesary now if you just use the html5 option of "Location for this client is static"?

Search Airheads
Showing results for 
Search instead for 
Did you mean: