eduzhai > Applied Sciences > Engineering >

Improving the Performance of Deep Learning for Wireless Localization

  • Save

... pages left unread,continue reading

Document pages: 8 pages

Abstract: Indoor localization systems are most commonly based on Received SignalStrength Indicator (RSSI) measurements of either WiFi or Bluetooth-Low-Energy(BLE) beacons. In such systems, the two most common techniques aretrilateration and fingerprinting, with the latter providing higher accuracy. Inthe fingerprinting technique, Deep Learning (DL) algorithms are often used topredict the location of the receiver based on the RSSI measurements of multiplebeacons received at the receiver. In this paper, we address two practicalissues with applying Deep Learning to wireless localization -- transfer ofsolution from one wireless environment to another emph{and} small size oflabelled data set. First, we apply automatic hyperparameter optimization to adeep neural network (DNN) system for indoor wireless localization, which makesthe system easy to port to new wireless environments. Second, we show how toaugment a typically small labelled data set using the unlabelled data set. Weobserved improved performance in DL by applying the two techniques.Additionally, all relevant code has been made freely available.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×