From RFID to IoT Bridging the last tens of meters RAIN RFID Meeting, June 2016 Vytas Kezys, Paul Slaby June, 2016
Outline Introduction - the need for a IoT device connectivity infrastructure RAIN - pro s and con s for IoT Technical limits and tradeoffs - coverage challenges Teslonix Vision - IoT device connectivity platform Location Security Where do we end up
Introduction GOAL: Affordably link billions of things interfacing to the physical world around us Need IoT device connectivity infrastructure: Connecting billions of things at the edge of the IoT network Power, locate, and communicate with all these things Get data up to the Cloud Could RAIN standard make scalable IoT deployments possible?
What s connected today? o Electronic devices such as: Nest thermostat, washer/dryer, electric vehicles What about things that are not devices? o A pair of jeans o Grocery items o Medication container o Issues with connecting things : o No own power source o No electronics to leverage o Low cost items Connecting things Challenge: Infrastructure tailored to lightweight nodes with ubiquity and cost effectiveness of Wi-Fi infrastructure Is RAIN a good fit?
RAIN might be the way Identify, locate, authenticate, and engage items in our everyday world RAIN advantages: Passive tags - ultimate in low-cost (vs BT LE, Wi-Fi etc) Only real contending standard for passive UHF tags UHF band is good for propagation and harmonized globally RAIN disadvantages: Passive limits range / coverage UHF band may not be best for locating tags RAIN s Place In The IoT Source: RAIN RFID
Ecosystem issues Tagged items - already coming for business and consumers Reader Infrastructure - not here for consumers Cost per area of coverage very high Needs to approach ubiquity of Wi-Fi infrastructure Source: LinkedIN
Technical limits and tradeoffs Regulatory limits on maximum range for ideal conditions Sensor tag Conventional tag High sensitivity tag Battery assisted tag Sensitivity -8 dbm -20 dbm -23 dbm -30 dbm Ideal Range 4 m 17 m 24 m 54 m
Ideal vs realistic Ranges were for ideal conditions (LOS, etc) Real deployments must consider: multipath fading shadowing tag desense tag orientation polarization etc With random effects: receive power at tag is distributed 25m x 25m room: 1000 tags distributed with random locations/orientation 4 readers at corners: max dist to tag is 17m Sensor tag Conventional tag High sensitivity tag Battery assisted tag Sensitivity -8 dbm -20 dbm -23 dbm -30 dbm Ideal Range 4 m 17 m 24 m 54 m Read rate 4 conventional readers 3% 32% 54% 88%
Market requirements Always-on and Always-connected for everyday things requires a robust and ubiquitous infrastructure o Not just handheld scanners/portals Cost of ubiquitous coverage Currently RAIN: >> $1 / sq ft Compare to Wi-Fi (2x $100 routers cover 3000 sq ft house): 7 / sq ft We have a long way to go!
Teslonix Vision IoT device connectivity platform Distributed mesh of Internet of Things Access Points (IoTAs) forming a tightly coordinated infrastructure Deliver more power on down link Provide better uplink sensitivity Distributed sensing of tag response to provide accurate and robust location Make scalable IoT deployments possible Achieve active device functionality using more affordable passive nodes Target: Infrastructure cost on par with high-end Wi-Fi Reader functionality built into Wi-Fi APs? High level integration to drive costs down
Same 1000 tags distributed with random locations/orientation in 25m x 25m room 4 readers at corners, max dist to tag is 17m Conventional readers; vs Tightly coordinated readers Example solution Sensor tag Conventional tag High sensitivity tag Battery assisted tag Sensitivity -8 dbm -20 dbm -23 dbm -30 dbm Range (ideal) 4 m 16 m 23 m 52 m Read rate 4 conventional readers Read rate 4 tightly coordinated readers 3% 32% 54% 88% 20% 88% 93% 99%
Can it work? Single reader - only one tag read Tight coordination - all tags read 6-8dB link budget improvement Stack of 10 jeans 3 tightly coordinated readers Each with a conventional RAIN tag, de-sensed due to tag-tag proximity
Location Location accuracy is use case dependent: 1 meter vs 30 cm Robustness to multipath is a problem With RAIN, 1 m accuracies are possible, but may require dense infrastructure deployment In early communications, multipath was seen as interference - a problem In current wireless, multipath is exploited: MIMO, SDMA Multipath should be properly exploited in location as well Tightly coordinated, multiple IoTA approach offers advantages in localization
Security End-end security energy and computation constrained end node Infrastructure generally has more energy and computational resources Shift burden of security to infrastructure? Move some Cloud capabilities to the readers proximity decreasing the latency reducing communication load decreasing processing load on the core network (better scaling)
Where do we end up Expected outcome IoT infrastructure with passive sensor/actuator nodes Highly integrated RF/digital SoC (System-on-Chip) for cost effective IoT infrastructure Novel silicon solutions for passive sensor nodes Impact New and more effective functionality in powering, locating, tracking and communicating with nodes Opens mass IoT deployments Enables smart applications An economic solution for massive deployments of billions of IoT nodes
QUESTIONS / COMMENTS
A better way to manage things (IoT) in our lives Confidential