Coexistence with WiFi for a Home Automation ZigBee product

Coexistence with WiFi for a Home Automation ZigBee product

Federico Dominguez, Abdellah Touhafi, Jelmer Tiete and Kris Steenhaut ,Erasmus Hogeschool Brussel

Vrije Universiteit Brussel,Brussels, Belgium

Communications and Vehicular Technology in the Benelux (SCVT), 2012 IEEE 19th Symposium on, 6-16 Nov. 2012

Adviser: Presenter: Yu-Jhang ChenDate: 102/12/18

Outline

1. Introduction2. Materials and Methods3. Results4. Discussion5. Conclusion

Introduction(1)

• Home Automation(HA):

1. Heating Ventilation Air Conditioning (HVAC) 2. automatic garage doors3. intruder detection alarms4. smart sensors detect temperature

Introduction(2)

Introduction(3)

• Tradition HA: – Expensive – Cabling difficulty

• Now HA:– Cheap– Smaller– Offer wireless connectivity

Introduction(4)

• HA products exploit frequencies bands :– 433MHz, 868MHz, 900MHz, 2.4GHz

• The 2.4GHz compared with 433MHz :– Higher data throughput– Worldwide availability as an Industrial Scientific

Medical (ISM) band

Introduction(5)

ZigBee:

• Advantage – low cost SoC platforms– High interoperability

• Defect– WiFi, Bluetooth, Microwaves, cordless phones interference

Materials and Methods(1)

Window Shutter HA system:• Content :

– Shutter motor and Remote Control (RC)

• Extra features :– Position and error feedback– Security– Scalability

Materials and Methods(2)

• The minimal Window Shutter HA system setup is just a Wall RC and a Window Shutter. The system can later be expanded with other RC models and more shutters

Materials and Methods(3)

ZigBee PRO:• Devices :– Texas Instruments(TI) CC253x SoC– 8051 8-bit microprocessor– AES coprocessor– USB controller– CC2591 RF

Materials and Methods(4)

Channel Allocation:• ZigBee defines 27 channels :–One in the 868 MHz band– Ten in the 915 MHz band– 16 in the 2.4 GHz band

• WiFi channels:– Range from 1 to 13

Materials and Methods(4)

• system uses ZigBee channel 26 and is vulnerable to interference from WiFi channel 13

Goals of tests :• Confirm or negate whether WiFi interference

poses a threat to the product• Find recommendations to avoid interference• Find a simple method to detect the presence

of disruptive interference during product installation

Materials and Methods(5)

• Packet Reception Rate (PRR)− Added a test function in the firmware− Counts all ZigBee packets flagged as test packets − RC have the capability to send test packets in

bursts of 1000 packets per second

Materials and Methods(6)

System Responsiveness (SR):• Assumed that a user expected− Response to UP, DOWN and STOP commands by

an RC , around 300ms − response of shutter position feedback in the

embedded LCD screen on the Multi RC while the shutter is moving

Materials and Methods(7)

• Cr is successfully executed commands • Cd is noticeable delay time• Ct is total number of commands • Fw is a weighted value for the quality of the

LCD visualization

Materials and Methods(8)

Materials and Methods(9)

WIFI interference:• Constructed a lab to test the affects:– Put in a metallic window frame to emulate– Used a Linksys WRT54GL and D-Link DIR-615– Used iperf to generate synthetic UDP test traffic

Materials and Methods(10)

• X: distance of interference source

• Y :WiFi traffic level • C:WiFi channel

Materials and Methods(11)

• Traffic rate Y equivalency in WiFi load

Materials and Methods(12)

• At distances below 5 meters from interfe-

rence source the sy- stem is practiuzally unreachable. The ef- fects of interference consistently dissipate after 15m in all chan- nels

Results(1)

• SR levels of 10% or less, where the sy-

stem is completely unresponsive,were observed at dista- nces of 5 meters or less from inter- ference source.

Results(2)

• Even at this rel- atively low traf- fic level a small but noticeable degradation (a- round 80%) of system respon- siveness was observed.

Results(3)

Confirm:• WiFi can create harmful interference to

ZigBee systems• Distance to interference source and WiFi

traffic level are key variables• Difficult to avoid the WiFi generate

interference with real traffic

Discussion(1)

• Distance from interference source– A distance of 5 meters can be tolerated– 15 meters seems to be a safe distance to avoid

harmful disruption

• PRR vs. SR– PRR can give a me accurate measure of the

expected performance– Measuring PRR is much simpler than measuring SR

Discussion(2)

• PRR vs. SR has a correlation value of 0.89. PRR% can the-

refore be used to predict the performance of the HA sys- tem.

Discussion(3)

• ZigBee can coexist with WiFi in a typical home environment

• Precautions are taken into account• New standards such as Wireless HD and

WiGig will not occupy the 2.4 GHz band

Conclusion(1)