Download to read offline
Uploaded byJorge Hidalgo
JavaOne 2015 - CON6489 - Smart Open Spaces Powered by Low Cost Computers
The document discusses the concept of smart open spaces, emphasizing their architecture, use cases, and real-time data tracking capabilities. It highlights how the use of Bluetooth LE and single board computers such as Raspberry Pi aids in the detection of devices and analysis of customer behavior in various environments, including stores and hospitals. The presentation also outlines the benefits of low-cost solutions and open standards for easier scalability and integration.
JavaOne 2015 - CON6489 - Smart Open Spaces Powered by Low Cost Computers
- 1. Copyright © 2015Accenture All rights reserved. Smart Open Spaces Powered by Java ME, Java SE and Single Board Computers Jorge Hidalgo & Julio Palma JavaOne Conference – CON6489 – October 2015
- 2. Copyright © 2015Accenture All rights reserved. 2 Presenter Introductions Jorge Hidalgo @_deors http://deors.wordpress.com Senior Technology Architect – Accenture Delivery Center in Spain Capability Lead – Custom Distributed & Architecture domain Father of two kids, husband, whistle player, video gamer, sci-fi junkie, Raspberry Pi fan, gadgets maniac... My other car is a Millenium Falcon. Julio Palma @restalion Technology Architect – Accenture Delivery Center in Spain Team Lead – Custom Distributed & Architecture domain Mountain biker, SW & LOTR fan, gamer, Nyarlathotep enemy, father of two kids who show me something new every day, husband. In my spare time I work at Accenture.
- 3. Copyright © 2015Accenture All rights reserved. 3 Objectives for the Session • Describe Open Spaces and common use cases • Introduce the proposed Architecture • Highlights of how the solution works • Examples • Live Demo
- 4. 4Copyright © 2015Accenture All rights reserved. What are Open Spaces and common use cases
- 5. Copyright © 2015Accenture All rights reserved. 5 Smart Open Spaces
- 6. Copyright © 2015Accenture All rights reserved. 6 Smart Open Spaces
- 7. Copyright © 2015Accenture All rights reserved. 7 Smart Open Spaces
- 8. Copyright © 2015Accenture All rights reserved. 8 Smart Open Spaces
- 9. Copyright © 2015Accenture All rights reserved. 9 Smart Open Spaces
- 10. Copyright © 2015Accenture All rights reserved. 10 Smart Open Spaces Department stores, smaller stores, hypermarkets
- 11. Copyright © 2015Accenture All rights reserved. 11 Smart Open Spaces Department stores, smaller stores, hypermarkets Museums, airports, train/bus stations
- 12. Copyright © 2015Accenture All rights reserved. 12 Smart Open Spaces Department stores, smaller stores, hypermarkets Museums, airports, train/bus stations Hospitals
- 13. Copyright © 2015Accenture All rights reserved. 13 Smart Open Spaces Department stores, smaller stores, hypermarkets Museums, airports, train/bus stations Manufacturing plants, oil rigs Hospitals
- 14. Copyright © 2015Accenture All rights reserved. 14 Smart Open Spaces Department stores, smaller stores, hypermarkets Museums, airports, train/bus stations Manufacturing plants, oil rigs Offices Hospitals
- 15. Copyright © 2015Accenture All rights reserved. 15 Smart Open Spaces Department stores, smaller stores, hypermarkets Museums, airports, train/bus stations Manufacturing plants, oil rigs Offices City Downtowns Hospitals
- 16. Copyright © 2015Accenture All rights reserved. 17 Smart Open Spaces I wish I could know what is exactly happening, real-time and historic info
- 17. Copyright © 2015Accenture All rights reserved. 18 Smart Open Spaces Smart in this context mean • Presence Zones platform to enable intelligent decision making • Real-time decisions • Strategy decisions
- 18. Copyright © 2015Accenture All rights reserved. 19 Smart Open Spaces Smart in this context mean • Track people activity through the radio signals of their personal devices • Smartphones • Wearables • Laptops • Tablets
- 19. Copyright © 2015Accenture All rights reserved. 20 Smart Open Spaces Smart in this context mean • Track people activity through the radio signals of their personal devices • Smartphones • Wearables • Laptops • Tablets
- 20. Copyright © 2015Accenture All rights reserved. 21 Smart Open Spaces Smart in this context mean • Track people activity through the radio signals of their personal devices • Smartphones • Wearables • Laptops • Tablets
- 21. Copyright © 2015Accenture All rights reserved. 22 Smart Open Spaces Smart in this context mean • Track people activity through the radio signals of their personal devices • Smartphones • Wearables • Laptops • Tablets
- 22. Copyright © 2015Accenture All rights reserved. 23 Smart Open Spaces • Now we can use collected information in real time: • Where are the customers located at this exact moment in time? • Which places are capturing the customer attention?
- 23. Copyright © 2015Accenture All rights reserved. 24 Smart Open Spaces • Now we can use collected information in real time: • Where are the customers located at this exact moment in time? • Which places are capturing the customer attention? • Or analyse aggregated data for insights on people habits • Which are the paths that customers follow more frequently inside the store? • A sale on selected items started yesterday at noon How was the activity in the surrounding area compared to normal days? • What is the activity pattern along the day in the electronics department? That information would be helpful to plan working shifts better
- 24. 27Copyright © 2015Accenture All rights reserved. Proposed Architecture
- 25. Copyright © 2015Accenture All rights reserved. 28 Smart Open Spaces Why develop a new solution for Presence Zones? “Vendor X already has a product named Y that does exactly this.” Our approach was born with one main objective TCO should be as low as possible To enable that objective - Leverage open standards, minimise cost of software licenses - Leverage low-cost, easy to obtain devices - Simple & lightweight, but easy to scale
- 26. Copyright © 2015Accenture All rights reserved. 29 Architectural Approach o Java and Python as programming languages/runtimes • Device sniffing through Java ME 8 midlet or Python script (for devices that cannot run ME... yet) • Data collector through Java SE 8 server o Bluetooth LE • More precise than WiFi • Not as frequently used compared with WiFi, but usage is growing fast thanks to wearables o Single Board Computers • Raspberry Pi A+, B+, 2 B • Beaglebone Black • Arduino
- 27. Copyright © 2015Accenture All rights reserved. 30 Architectural Approach Sniff devices on spot Send data to collector Correlate and store Generate reports (daily, on demand, real-time)
- 28. Copyright © 2015Accenture All rights reserved. 31 Architectural Approach Device (Nexus 5) detected by RPis & Device (Pebble) detected by RPi Device (iPhone 6) detected by RPi Four packets sent to collector Matched by MAC address and timeframe Only three events are generated
- 29. Copyright © 2015Accenture All rights reserved. 32 Architectural Approach Device (Nexus 5) detected by RPis & After some time, detected only by RPi After some time, detected only by RPi Events are timestamped Easy to obtain time series for a given device
- 30. Copyright © 2015Accenture All rights reserved. 33 Architectural Approach – Scaling Out Data Collector Edge Devices Data Centre / Cloud Analytics Historic Info Real-Time Info Sockets Sockets Sockets MQTT Internet DomainIntranet Domain
- 31. Copyright © 2015Accenture All rights reserved. 34 Architectural Approach – Scaling Out
- 32. Copyright © 2015Accenture All rights reserved. 35 Architectural Approach – Scaling Out acme/activity/europe/uk/edinburgh/electronics
- 33. 36Copyright © 2015Accenture All rights reserved. Examples and Demo
- 34. Copyright © 2015Accenture All rights reserved. 37 Examples and Demo – Reports Distribution of time at beacon per device Average time per beacon
- 35. Copyright © 2015Accenture All rights reserved. 38 Examples and Demo – Reports Frequent Paths
- 36. Copyright © 2015Accenture All rights reserved. 39 Examples and Demo – 3D Real-Time
- 37. Copyright © 2015Accenture All rights reserved. 40 Examples and Demo – Hot Zones
- 38. 41Copyright © 2015Accenture All rights reserved. Summary
- 39. Copyright © 2015Accenture All rights reserved. 42 Conclusion – Lessons Learned Raspberry Pis are cheap but powerful enough to take multiple roles simultaneously: Bluetooth device detection Collector device correlating data and producing reports When multiple languages can do the job, leverage the skills of your team More complex platforms like OEP are very useful... ...but can be overkill – mind the KISS principle Open standards facilitate integration and addition of new features in the future
- 40. Copyright © 2015Accenture All rights reserved. 43 If you want to get in touch Public Accenture Portal Sub-site for Application Services for Java: https://www.accenture.com/us-en/service-enterprise-systems-java.aspx Follow (and interact with) us on Twitter: https://twitter.com/AccentureTech https://twitter.com/_deors https://twitter.com/restalion
Editor's Notes
- #2 Hello, Wellcome everybody to our conference sesión six thousand four hundred eighty nine, Smart Open Spaces.
- #3 My name is Julio Palma, I’m Technology Architect from Accenture based in Málaga, in Spain.
- #6 To help us to visualize this problem I need help from a friend of mine. Let me introduce you Mark. He's very excited in his new role as an open space manager.
- #7 As an Open Space Manager he needs to manage spaces with a little number of customers …
- #8 ... and others with a huge amount of them.
- #9 Customers can be indoors ...
- #10 ... or outdoors
- #11 He is not restricted to one specific industry, he can manage Department stores, smaller stores, hypermarkets
- #12 Museums, airports, train/bus stations
- #13 City downtowns, hospitals
- #14 Manufacturing plants, oil rigs
- #15 and, of course, offices
- #16 and, of course, offices
- #18 Mark wants to implement a “Smart” aproach to his new role. We can imagine a scenario where our common friend is facing a simple (aparentely) problem. How can he make this space Smart?
- #19 One of the first things we can imagine is to create some “Presence Zones” within we are going to monitor people activity.
- #20 In order to be as unintrusive as possible we use signals of their personal devices: Smart phones, weareables, laptops, tablets…
- #21 When customers are into the “presence zones” this information is captured and retrieved in different times
- #22 , we are taking “shots” of where each device is at this exact moment in time.
- #23 In order to have a look that let us know, let us… understand, how customers are ‘moving’ into our “Open Space”
- #24 We can use collected information in multiple ways: We can use it in “real time” trying to answer questions: Where are the customers located at this exact moment in time? Or … Which places are capturing the customer attention? Which places has a high density of customer.
- #25 We also can use this information to analyse aggregated data, in this case we use this tool as a way to understand customer habits: which path customers follow inside the store, where they spent more time (special offers?, at electronic or music department?), which places in the store, hospital, airport… are capturing people’s attention. All this information helps us to plan, for example, working shifts of attendants or space distribution, better.
- #29 Why do we develop a new solution for presence zones? We know that there is previous products that does exactly what we wont. This aproach has one main objective: TCO should be as low as possible. - we’re going to use open standards, minimise the cost of software licenses. we’re basing the solution on low-cost devices, and we want to use easy to obtain devices We want a simple and lightweight solution, and of course, easy to scale. -
- #30 For the technology selection we indentify Python and Java as the programming languages and runtimes that we wanted to use for the platform. Basically we’ll have a device sniffing tool that uses a Java ME midlet or Python script to sniff all the devices arround the area (know about their ssignal strength, mac adress,…), sending all this information to a data collector. The data collector, also can run on a python script or a Java SE server application, we are going to integrate all the different sniffers, and all this collected information in just one collector platform. To listen to he devices we’re using Bluetooth Low Energy technology because is more precise than WiFi, is nor as frequently use as WiFi but its usage is growing fast, mostly thanks to wearables. All this software runs on single board computers such as Raspberry Pi, Beaglebone or Arduino. In our life demo we’re going to use Rapberry Pi