Cellular Positioning Shashika Biyanwila Research Engineer.

Cellular PositioningShashika Biyanwila

Research Engineer

Cellular Positioning

Outline

• What is Cellular Positioning• Positioning Parameters• Feasible Approaches Identified• Implementation• Some Trial Results• Future Approach of the Research

Cellular Positioning

What is cellular positioning ?

• Determining the position of a Mobile Station (MS), using location sensitive parameters

• Why ????– To provide Location Based Services…

Cellular Positioning

Operator services

BillingNetwork management Location based services Wireless Gaming

Assistance

Roadside assistancePersonal or vehicle emergency Alarm managementDriving DirectionsTracking

Tracking criminalsTracking external resources containers etc

Monitoring

Monitoring delivery processFleet & freight trackingPersonal Child SecurityMobile Worker management

Information

TrafficNearest servicenewsnavigation helpadvertisingInformation Directory

Applications of cellular positioning

Cellular Positioning

• Cell-ID• Received Signal Strength Intensity (RSSI)• Timing Advance (TA)• Uplink Time (Difference) Of Arrival (TDOA)• Downlink Observed Time Differences (E-OTD)• Angle of Arrival (AOA)

Positioning Parameters

Cellular Positioning

Feasible Approaches Identified

Positioning Techniques

1Geometrical

Approach

2Statistical Approach

3Database CorrelationApproach

Cellular Positioning

1. Geometrical Approach

• Based on distance measurements

• Two Steps:

- Distance calculation

- Location Estimation

Cellular Positioning

Geometrical approach contd..

• Distance Calculation - Measure the RSSI from neighboring cells - Apply Propagation models to calculate the distance

• Propagation Models- Hata Model- Extended Hata Model- Lee’s Model- CCIR Model- Walfisch-Ikegami Model (for micro cells)

Cellular Positioning

2. Statistical Approach

• Construct a statistical propagation model for the RSSI- Find RSSI at distance d from the transmitter- Offsite calibration is necessary to estimate the

propagation parameters

• Define a probability distribution for the RSSI• Location estimation problem is solved as an inverse

or, rather, inference problem

Cellular Positioning

Statistical Approach contd..

• Log-loss or Log-distance model • Gaussian Probability Distribution• Propagation Parameter Estimation

- Maximum Likelihood estimation• Location Estimation

- Maximum A posteriori Probability

Cellular Positioning

Statistical Approach cntd..

• Area being considered is divided into several squares

• A posterior probability of the location be within a square, is calculated for each square

Square with Maximum A posterior

Probability

Cellular Positioning

3. Database correlation Method (DCM)

• Involves a database of reference fingerprints for the whole area of interest.

• Fingerprint – a recorded measurement sample from a certain location in the area

GPS coordinates of a location

RSSI (from available cells) in that location

Cellular Positioning

• How to collect fingerprints?• By measurements• Using a Network

planning tool

• High sampling resolution is needed.

Measurement

Fingerprint

Test route

Fingerprint

DCM contd…

Cellular Positioning

• Location estimation

•Compare the input measurement with reference fingerprints

- Using Cost Functions•Location of the best matching reference fingerprint

Estimated Location

Input Measurement

DCM Algorithm

Database

Estimated Location

DCM contd…

Cellular Positioning

Implementation

RSS

Measurement Unit

RSSI + GPS Reading

Commands•Interfacing Program•Database•Location Estimation Algorithm•Display Program

Software environment

Location ?

Hardware Environment

Cellular Positioning

Trial & Results

• Urban

- Wellawaththa to Kolpetty

• Suburban

- Katubedda to Piliyandala

• Rural

- Ibbagamuwa

Cellular Positioning

Urban area…..

CDF wise comparison for Urban area

0.00%

20.00%

40.00%

60.00%

80.00%

100.00%

120.00%

100 200 300 400 500 600 700 800 900 1000

Error Less Than (m)

Per

cent

age Geometrical

Statistical

DCM

Existing Method

Cellular Positioning

Suburban area ……….

CDF wise comparison for Sub Urban area

0.00%

20.00%

40.00%

60.00%

80.00%

100.00%

120.00%

100 200 300 400 500 600 700 800 900 1000 1500 2000 2500

Error Less than (m)

Per

cent

age Geometrical

DCM

Existing Method

Statistical

Cellular Positioning

Rural area ……..

CDF wise comparison for Rural area

0.00%

20.00%

40.00%

60.00%

80.00%

100.00%

120.00%

1000 1250 1500 1750 2000 2250 2500 2750 3000 3500 4000 4500 5000

Error Less Than (m)

Per

cent

age

Geometrical

DCM

Existing Method

Statistical

Cellular Positioning

Future Approach of the Research

Improvements to the current DCM approach• Drawbacks

- Few instances of poor estimations

- Creating, updating & maintenance of the database

• How To Overcome

- Refined estimation techniques

- Use of a Network planning tool to create fingerprints

Cellular Positioning

Implementation of a positioning engine and associated services

Services• Get your own location • Track others – web-based location on a map

GSM Network

Estimated Location

Received Signal

Fingerprint

Location Estimation using Received Signal Fingerprint & database

System

Information

Calibration

Fingerprints

Digital MapsPositioning Engine

Thank You