Skip to main content

DMAP Summary

·2 mins

Goals #

  • to establish the accuracy of Mist indoor location
  • to define impediments to implementation of Mist systems for location tracking
  • to propose ways to improve Mist indoor location accuracy

Tests #

The Mist Devices were Tested in Three different Configurations #

  1. Straight-line

  • low accuracy. Straight-line configuration does not work very well. It may be good enough for hallways where there is some interaction between the APs in the hallway, and those in surrounding rooms.
  1. Single-room

  • Much better accuracy than the straight-line configuration, but the client moves around quite a bit even when still. It would seem this is a result of signal interference between the APs because they are too close together.
  1. Full-building

  • The best accuracy of all the tests. The location smoothing seems to be less erratic when the APs are spaced out, but the latency issue still remains.

Naecon Demo #

  • for the Naecon demonstration three Mist devices were installed, and the data from Bluetooth clients were extracted using the Mist REST API.

Summary Findings #

  • accuracy of the Mist system is within 1m
  • latency for the location information is typically around 2 seconds
  • the Mist APs perform sporadically through walls so each open area should be treated as it’s own installation
  • too many or too few APs will affect performance negatively
  • 1 AP per 1500 ft2 room-like area
  • APs separated by about 55 ft in hallways or passageways

Video #

The video linked below shows the promise, and main issues with the MIST system. In the video what you see (from left to right) is the Mist Dashboard showing SDK Clients, This same information being displayed on an OpenStreetMaps map as was done for Naecon, and a synchronized video showing the SDK client’s actual position within the building.