Some optimizations of WiFi- Based Indoor Positioning Hanqing Liu - - PowerPoint PPT Presentation

Hanqing Liu

Some optimizations of WiFi- Based Indoor Positioning

CONTENTS

Research Background

01

Related Work

02

Our Project

03

Follow-up Work

04

Research Background

Research Background

4 The implement of positioning in indoor environment.

I ndoor localization W HY?

The signal from the satellite is too weak.

How ?

Using wireless communication, base station positioning, inertial navigation to form indoor location positioning system

Related Work

Basic Concept

6 The basis of Wifi-based indoor positioning system! Can be predicted by a LDPL m odel.

Received Signal Strength I ndication( RSSI ) LDPL Model

If Pi and γi are known, then an RSS m easurem ent pij can be converted into the distance dij.

A Basic Positioning Method

7

Triangular positioning and LDPL Model

The LDPL Model provide a connection between the location of the user and RSS. The cross point does not always occur in real case!

Some Advanced Scheme of Indoor Positioning

8

Fingerprint-based positioning

 Offline phase: a site survey is conducted to collect the vectors of RSSI.  Online phase: a user samples or measures an RSSI vector at his/ her position and reports it to the server.

Some Advanced Scheme of Indoor Positioning

9

Peer Assisted Positioning

 The device broadcasts a special audio signal.  Distance between peers is calculate based on TOA.

SLAM

 The training phase  The operating phase

Our Project

1 1

Our indoor positioning based on RSS

1 2

Stage 1: Triangular Positioning Only

1 3

Stage 2: Triangular Positioning With Some Optimizations

1 .Estim ating position based on position history. 2 . Estim ating RSS value through Kalm an filtering

Prediction = last_est + kg* (rssi- last_est)

Follow-up Work

1 5

Follow-up work of project

1 .Pressure test 2 .Adjustm ent of API

W HY? To find Bottleneck!

1 6

Pressure Test Bottleneck: 1 .Connection to DB 2 .Max Connection to Apache 3 . CPU

THANKS