DYNAMIC BLUETOOTH BEACONS FOR PEOPLE WITH DISABILITIES A journey from ibeacon to IoT beacons, InfinIT Summit 2017
BLUETOOTH BEACONS Short information sent by radio A few times per second Kind of radio barcode Similar to active RFID but using consumer technology Bluetooth beacons ibeacon is Apple s subset (2013) Eddystone is Google s variant Physical Web (2014) Custom variants Can be deployed in different setups: Wikipedia 2017-09-06 Side 4
SETUP 1) FIXED BLUETOOTH BEACONS Most common setup Radius Networks 2017-09-06 Side 5
SETUP 2) MOBILE BLUETOOTH BEACONS Useful for automatic triggers Radius Networks 2017-09-06 Side 6
ABILITIES OF A BLE BEACON What it can Provide a vague approximation of distance Broadcast a small amount of information May work without Internet connectivity ~10cm to ~30m What it cannot Cannot provide accurate direction (Except with some workarounds) Cannot detect smartphones or other Bluetooth devices Limiting factors Battery time Emitting is cheap (by the beacon) Scanning is expensive (by the sensor / smartphone) Scaling IoT security 2017-09-06 Side 7 ABCDEFGHIJKLMNOPQRSTUVWXYZ012345
BLUETOOTH BEACONS COMPETITORS 2D-barcodes Much cheaper (if using paper), but requires an active start by the user only for line-of-sight less appropriate for nearby discoveries NFC Cheaper, safer, but only for very short range RFID Bluetooth beacons are a kind of active RFID Traditional RFID cannot interact with smartphones Geomagnetic positioning (compass) Requires a pre-mapping and maintenance Wi-Fi sniffing Lower temporal precision With / without trilateration 2017-09-06 Side 8
COMPETITOR: WI-FI POSITIONING 2017-09-06 Side 9
IBEACON TECHNOLOGY Apple proprietary convention Subset of BLE (Bluetooth Low Energy) Broadcast a (cryptic) number Data (up to 31 bytes) ibeacon prefix (9 bytes) Proximity UUID (16 bytes) Major (2 bytes) Minor (2 bytes) TX power (1 byte) 2017-09-06 Side 10
EDDYSTONE TECHNOLOGY Google s open beacon format: Physical Web Broadcast e.g. a short compressed URL https://goo.gl/aq18zf Does not necessarily require a dedicated app only some HTML E.g. with Google Chrome, Opera: <head> <meta charset="utf-8" /> <meta name="description" content=" 30 til Bygdøy via Bygdøynes om 2 min 56 til Solemskogen om 4 min 30 til Bygdøy via Bygdøynes om 12 min 30 til Nydalen klokka 14:54 56 til Solemskogen klokka 14:58" /> <meta http-equiv="refresh" content="30" /> <link rel="icon" href="ruter.png" /> <title>gullhaugveien</title> </head> Opera 2017-09-06 Side 11
CUSTOM APPROACH Add custom Bluetooth beacons to home/building devices Able to broadcast: a type: lamp, fan, music, TV, door, coffee machine a status: e.g. on / off, alarm, temperature, defect, toilets in use Able to actuate: e.g. switch on / switch off By accept incoming connections (GATT) Link to existing electronics Relays (to switch on/off any electrical device) Radio (e.g. 433 MHz, ZigBee, Z-wave for smart-house products) Infrared (for televisions) I 2 C 2017-09-06 Side 12
IMPLEMENTATION 1: WITH RFDUINO Microcontroller compatible with Arduino Easy custom electronics Built-in Bluetooth Suitable for running on batteries Examples: Control a 230V electric plug Via a 433MHz module 12V relay to control a lamp 2017-09-06 Side 16
IMPLEMENTATION 2: WITH RASPBERRY PI Full desktop Linux Built-in Bluetooth (Raspberry Pi 3) or with a dongle Better for multimedia and Internet connectivity Easier software prototyping Examples: Audio beacon To play a sound / speech synthesis / music, when needed Control infrared devices (e.g. TV) With infrared emitter (can also be done with RFduino) 2017-09-06 Side 17
IOT SECURITY CHARACTERISTICS Share all problems found in classic IT projects, and also: Physical integrity Operate over larger / public areas More controversial Hacktivism Security is expensive, also in energy Hardware costs, performance impact, energy impact Small sensors are not powerful enough for crypto IoT projects involve many different technologies Weakest link? IoT projects involve many different actors Trust? Distributed responsibilities 2017-09-06 Side 19
BEACONS COMMON PITFALLS Security Physical tempering: (re)move beacon Cloning: make the app believe it is elsewhere Piggybacking: competitor app Use some counter-measures Bluetooth pairing (only for small scale) Ephemeral identifier (Eddystone) App logic (geofencing verification, GATT connection ) Proprietary cloud (randomisation ) Scaling Push vs. scan Push: performance issues with ~2 to ~15s on ios, but performance drop on telephones with many beacon-using apps Scan: high battery consumption, problematic for background apps Maintenance 2017-09-06 Side 20
FINDINGS Dynamic Bluetooth Beacons Relatively cheap and easy to work with Versatile: can be adapted to many IoT items Work out-of-the-box with smartphones/tablets Robust, with fast reaction time More than 1 full interaction per second: Discovery, read status, change status, get confirmation Good spatial precision A distance of ~1 meter between beacons can be enough: If the distance between the tablet/telephone and a beacon is significantly shorter than the distance between two beacons Main limitations to gain adoption: Need to improve standardisation, maintainability, and security 2017-09-06 Side 22
DYNAMIC BEACONS: FUTURE DEVELOPMENT Take better advantage of: Google Eddystone open beacon format E.g. Physical Web, broadcasting a short URL For more standardisation, more scalability, security Together with Web Bluetooth API To avoid having to install apps Bluetooth 5 Longer range, richer payloads Emerging beacon-based standards 2017-09-06 Side 23
QUESTIONS? alexandre.alapetite@alexandra.dk 2017-09-06 Side 24