Indoor Wireless Localization

Name: Ali Reda Salim

In Collaboration with: Mohmamad Assaad, Daylight Katengwe

Supervisor: Zihuai Lin

Presentation Outline

•What is Indoor Wireless Localization and its applications •Previous research & bottlenecks •Our system design & approach •Analysis & Results •Demonstration Video •Conclusion & Future Work

Indoor Wireless localization

•GPS is a hug success and generated billions of dollars in revenues. •Worked seamlessly within line of sight (outdoor)•GPS signals can’t penetrate walls and solid concrete•The evolution of low cost wireless sensors permitted solutions for indoor positioning.• low cost 802.15.4 Zigbee compliant nodes to implement our proposed design.

Indoor Wireless localization

Applications •Tracking local assets •Medical applications •Internet of things •Emergency rescue operations •Commercial & business uses •Location based services

Indoor Wireless localization

•Previous papers used RSSI claim an accuracy of 1-1.5m•Other researchers relied on TOF & TDOA and claim an accuracy of almost 2m.•Above algorithms rely on RSSI, time stamps and node synchronization •Most papers used Zigbee compliant stacks and extracted the required metrics from network and application layers

TIMAC Protocol Stack • OSAL & HAL layers to access

lower layers • HAL layer provides a broad

range services

Z-Stack Protocol Stack • Z-Stack provides access to

upper layers• ZMAC Interface to access

MAC layer

Approach

802.15.4 MAC API• Initialization• Data• Management• Call-back

Approach

HAL API• Provides functions to access

the hardware services like timers, UART, and ADC.

• Separates the hardware from the software specific

LCD Service  HALLCDINIT() Initiate LCD screen on-board HALLCDWRITESTRING() Write strings to the screenHALLCDWRITESTRINGVALUE()

Write strings and values to the screen

KEY Service  HALKEYINIT() Initiate Key & Joystick on-board HALKEYCONFIG() Configure Keys operation Timer Service  HalTimerInit () Initiate timer services on-board HalTimerConfig() Configure and pre-scale timers halTimerIntEnable();  UART Service  HALUARTINIT () Initiate Serial port on-board HALUARTOPEN () Open a serial portHALUARTWRITE () Write strings and values to

serial port

Function/Event Description Prototype MAC_CbackEvent ( )

Call back function sends MAC events to the application layer

void MAC_CbackEvent(macCbackEvent_t *pData)

MAC_MCPS_DATA_IND

An event called by the call back function and sends data to app

macDataInd_t *pDataInd

MAC_MCPS_DATA_CNF

An event that sends data to app each time the MAC_McpsDataReq function is called

macMcpsDataCnf_t *pDataCnf

System Design

• TI CC2530 development kit

• 802.15.4, HAL & OSAL API’s

• TIMAC Sample Application

• Modification of TIMAC protocol stack to support timer services

• Integrating few Z-stack functions into MAC Sample application

System Design

• Hardware timestamp stored in 40-bit register upon transmit/receive of frames

• Average signal strength measured by reading the 8 symbols after the SFD domain

• RSSI and Correlation values • Data output to I/O devices

Analysis & Results

• Point to point network• Frame transmission Interval set to 5 seconds • Transmit and received frames sent to serial port• Clock drifts calculated between transmission

and receive time of various frames

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49175

180

185

190

195

200

205

RSSI Readings(Line of Sight vs Non Line of Sight)

C23 D22

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46150

155

160

165

170

175

RSSI Mid Range (Non line of Sight)

Series1 Series2 Series3

1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 910

50

100

150

200

250

300

Sci-Tech Libary Readings(Mobile Node)

RSSI Corr

Analysis & Results

Analysis & Results

• 3 static node & 1 mobile node

• RSSI values collected from each node & sent to serial port

• MATLAB simulations triggered by the serial input

• Estimated mobile node location within an accuracy of 1-1.5m

Video Demonstration

Conclusion

•Integration of Z-Stack & TIMAC software protocols.•A new approach to provide reliable & accurate metrics to any localization algorithm •Access of lower layers•Low Local clock speeds•Synchronization problems •Reliable RSSI & Correlation readings •Accuracy of almost 1m using just signal strength and correlation values

Future Work

•Implement smoothing techniques on local clock to minimize the effect of clock drift • Linear smoothing using linear regression to align local clock readings between 2 different nodes•Use external high speed clocks or nodes with higher local clocks •Synchronization of nodes wirelessly •Use antenna diversity to have more accurate RSSI readings •Ultra-wide band can be used instead of narrow band signals like Zigbee (5Mhz)

16

Thanks for Listening