Upload
jim-boland
View
162
Download
0
Embed Size (px)
Citation preview
Internet of ThingsAnd other interesting technology trends
Jim BolandSoftware Architect
Cognos AnalyticsIBM Canada Ltd
@neoslimjimhttps://www.slideshare.net/JimBoland3/hacking-health-iot-analytics-and-other-trends
Trends in computing
• Wearable technology
• iBeacons and micro-location context
• Contextual and anticipatory computing
• IoT, big data and analytics
• Wrap up: how all this is about plugging us into the IoT!
Mainframe Desktop PC Laptop Mobile (smartPhone/tablet)
Trends: Moore's law, computing distribution, personal/specialized, mobility ...
Wearables
vandrico.com/wearables
• Head: 63. 84
• Neck: 12 12• Chest: 10
21• Torso: 19
27• Arm: 8 11• Wrist: 119
213• Hand: 5 8• Fingers: 10
11• Legs: 9
13• Feet: 9
12
2014 2017
Wearables • Wearables complement (rather than replace) smartPhones• Wearable technology is about the more seamless integration
of tech:• Notifications - SmartWatches, SmartGlasses• Sensors - Fitness/health devices• Remotes and Identification • Virtual Assistants
SmartWatch & Glasses
• While much of the hype is with these, not the early success...
Fitness/Health
• Fitness/health devices the early leader in wearables
• Sensors to monitor heart rate, activity, sleep, etc.
• Nike abandoned the Fuelband
Remotes
• Want an interface like Minority Report and Tony Stark?
• Map physical gestures to intent, from:
• Fingers - Fin ring
• Arm - Myo ( Thalmic Labs)
• Wrist - Glance (kiwi wearables)
IdentificationWearables represent your identity for:
• Payment systems - Apple Watch and ApplePay
• Application/device - vivalnk tatoos
• Physical security - Kevo locks
• Contextual/anticipatory computing
As well, wearable biometric sensors create ways for establishing identity:
• Touch ID - Fingerprint scanners
• Nymi - heartbeat signature
Smart Shoes?
• Vibration in respective shoe to indicate right/left turn
Summary on wearables
• Overall themes: • Collect sensor data to gain context of the user• Unobtrusive means to feed information back to the user• Represent the user remotely (intent and identity)
• There are DYI prototyping kits out there (e.g. MbientLab’s Metawear)
• There are technical challenges still: security and power
iBeacon
• Conceptually similar to a GPS satellite, but at a micro-location (and often indoors) scale
• In a store with a beacon in each department, the phone app knows where it is, by which beacon(s) it sees
• iBeacons don't track/collect data from phones. Without an installed app, there is no interaction with the phone/beacon
iBeacons"I'm
beacon #53"
Micro-location
• Emergency ward patient tracking
• Contextual retail assistance
• Indoor mapping/directions
• Home automation
• Tealeaf/Google Analytics style analytics of in-store foot traffic
• Micro-location based geo-fencing/context
Contextual Computing
Better understand the context of a user request, in terms of:
• Context from device sensors
• Context from user interaction history
• Context from user profile/social graph
Providing context of who, where, when and the previous conversation
Siri example
Next request in context of first
Affective Computing
• Special case: using context of a user, to understand how people feel and emotions (and adapting responses as a result)
• With the pervasive sensors of smartphones and wearables, there is a new source of information
Anticipatory Computing
• Contextual computing: context is used to interpret a user request
• Anticipatory computing: monitors context and anticipates user's need without an explicit request
"Invisible buttons"• Amber Case [ESRI/Geoloqi] coined the term "invisible
buttons"
• Concept: instead of touching a button on a smartphone as the trigger for an action, the user becomes the trigger by entering a geofenced area
• Ultimate example of the User Interface getting out of the way
• Example: my house! (Kevo lock, 21 iBeacons, Hue lights, airPlay speakers)
IoT and Big Data
• IoT = devices sending and receiving data• IoT = LOTS of devices, sending LOTS of data, ALL
the time• IoT = Big Data!
1001
010
1001010
1001
010
1001
010
1001010
1001010
Making sense of big data
• How do you get value from all that data?• Analytics can help!
Data in motion
• Streaming analytics• Scalable solutions to tame the
firehose• Latency sensitive response• Realtime feedback• Moving averages• Identify out of range values• Smooth out jittery sensor data
Data at Rest• Batch analytics
• Historical analysis• Trends• Outliers• Correlations in the data• Predictive analytics
• Stream analytics, batch analytics - both have a role in IoT!
Database
Example IoT Architecture
Local Hub/IoT
Gateway
Device
Sensor
Sensor
Device
Bluemix
- Streaming Analytics- Batch Analytics- Other business logic
MQTTBLE
WIFI
ZigBee
Analytics in bluemix
• Trying analytics out!• Freemium account:
https://console.ng.bluemix.net
Watson Analytics
• Cognitive computing - machines doing the analysis
• Freemium account:
www.ibm.com/analytics/watson-analytics/
Node-Red
• https://nodered.org/ • Run locally (e.g. Rpi) or cloud (e.g Bluemix)