Background Image

Blog Post

Mar 02

Cisco Hyperlocation Real World Testing


Cisco Hyperlocation Real World Testing

We've recently been doing some testing with Hyperlocation, with the aim of seeing what the technology can do in the real world, not just in a pre-orchestrated demo. The results are quite promising, proving that while not the cheapest option for accurate location, the Hyperlocation Module lives up to its promises.

The photograph of the unit is shown on the right. Its an antenna array, which sits over the top of the AP, plugging into the WiPS unit (installed in the back of the AP). You're looking at the Cisco 3700 series AP in this photo, which is one AP which can take that unit.

 

a

The sample test area is illustrated below, consisting of an open environment, with the test locations being the desks where people would be typically sitting. Note that there was a rather heavy set of metal framed glass doors between the rooms. 

Hyperlocation Accuracy: Results

The table below shows the accuracy levels achieved in each of the sample locations. The unit achieved sub 10m 100% of the time, which is a good start. The average error was at worst under 4m and at best a little over a half metre - pretty good values.

Test Name Location Computation Frequency(s) 10m Accuracy (%) Average Error Distance (m) 90% Accuracy Error Distance (m) 75% Accuracy Error Distance (m) 50% Accuracy Error Distance (m)
Location 1 2.7 100 1.11 1.67 1.33 1.01
Location 2 2.8 100 0.63 1.47 0.93 0.42
Location 3 2.7 100 2.59 4.71 3.41 2.28
Location 4 2.7 100 1.39 2.01 1.8 1.31
Location 5 2.6 100 2.05 3 2.65 1.82
Location 6 2.6 100 3.92 5.15 4.96 4.38

 

Hyperlocation Accuracy: Explanation

You can see that the average error distance is easily sub 3m, with test # 6 being the worst result. The reason for test #6 being less accurate is due to the fact that it is placed outside of the “triangle” coverage area of the APs meaning that there is less calculation data to provide accurate results. This is important to note as there will always be scenarios where a client will fall outside of an optimal AP placement area leading to less accurate calculations.

 

The last important factor to note is about client roam sensitivity in relation to accurate location results. If a client is connected to the closest AP possible (optimal) then it will be in the best position to produce accurate results. Say for instance if the client was “sticky” and was associated to an AP 15 meters away even though there is a better (closer) choice, then the location accuracy will be reduced due to sub optimal calculations from the distant hyperlocation AP. This is why proper RF design is paramount to optimizing accuracy.

 

The CMX Phone App

The CMX phone app is a really handy way to check the accuracy - screen shot shown below.

 

 

 

Hyperlocation: Conclusion

With a proper RTLS RF design, the Cisco Hyperlocation modules work extremely well. Allowing sub 3m accuracy's when within the "triangle" coverage areas of the access points.

Note that there's a few lessons learnt:

  • Accuracy of placement of APs in CMX is well, hyper important
  • AP height and orientation should be recorded
  • Keep the APs on the straight and level - don't tilt!

If you are interested in our RTLS designs with Hyperlocation, visit our Wireless Networks Page, or email sales@iptel.com.au

Back   
 
Add Comment:
Please login or register to add your comment or get notified when a comment is added.