Compressive Sensing. ⢠Scalability and Energy Efficiency ... Power of the IoT+CrowdSensing ... On phone, on broker (SenseDroid, SATWARE). ⢠Techniques ...
Essence: A Machine Learning Approach to Scalable and Energy-‐Efficient Sense-‐making for Internet-‐of-‐Things (IoT)
Santanu Sarma University of California Irvine September 2015
Overview • • • • • •
Sense-‐making in IoT Problem Descrip9on Compressive Sensing Scalability and Energy Efficiency A Middleware Perspec9ve Summary
IoT Symposium, Oct 2015
© S.Sarma
2
Internet-‐of-‐Things
IoT Symposium, Oct 2015
© S.Sarma
3
Internet-‐of-‐Things
[Vermesan2011]
IoT Symposium, Oct 2015
© S.Sarma
4
IoT Trends
IoT Symposium, Oct 2015
© S.Sarma
5
What is Sense-‐Making? • An informa9on-‐theore9c perspec9ve to mean the process of understanding the connec1ons (among people, places, environment, objects, and events) – in order to an9cipate their trajectories – develop meaningful insights – act effec9vely using empirical data /insights
IoT Symposium, Oct 2015
© S.Sarma
6
IoT+CrowdSensing
Pushing toward more interven9on
7
IoT Symposium, Oct 2015
© S.Sarma
Power of the IoT+CrowdSensing • Using mobile crowdsensing to – Leverage already deployed smartphones – Extend the ranges of exis9ng in-‐situ sensors – Send mobile users to specific loca9ons
• Crowdsensing broad use cases – Disaster and emergency response – Personal health monitoring and wellness – Smart spaces and their effec9ve u9liza9on IoT Symposium, Oct 2015
© S.Sarma
[Yuen2011] 8
Sensing -‐> Sense-‐making Severity Alert System
Personal Sensing to indicate Fall detec9ons, injury severity, alerts in old age people to provide scalable health care IoT Symposium, Oct 2015
© S.Sarma
9
Sensing -‐> Sense-‐making Radia9on field near Fukushima
Hazardous gas in campus
Spa9al Field Sensing With Mobile Sensors
IoT Symposium, ct 2015 Latest Informa9on Crisis Map SOhowing
© S.Sarma
10
Sensing -‐> Sense-‐making • Avoiding congested streets in a city • Finding the most popular booth in a fair • Searching for the ride with shortest lineup in an amusement park
IoT Symposium, Oct 2015
© S.Sarma
11
Sense-‐Making : Purpose & Goals u A Framework for Sense-‐making using mobile phone & infrastructure sensors to derive insights u Powerful addi9onal sensing abili9es and features for community of users by community of users u Understand user and group context efficiently u Develop a efficient mechanism (for energy , performance) IoT Symposium, Oct 2015
© S.Sarma
12
The Problem – A cross layer, end to end issue § Several barriers and huge investment of 9me to build sense-‐making applica9ons § Lack of a framework hinders ease and speed the development of sense-‐making apps § Non-‐Scalable, Ad-‐hoc, non-‐standardized API § Unsupported network infrastructure, and configura9ons
IoT Symposium, Oct 2015
© S.Sarma
13
Middleware Plalorms and Techniques for Sense-‐making • On phone, on broker (SenseDroid, SATWARE) • Techniques implemented in middleware – Compressive and Collabora9ve Sensing – Virtual Sensing for Sense-‐making – Seman9cs Driven Sensing and Actua9on
• Combining In-‐situ Sensors with Mobile Crowdsensing
IoT Symposium, Oct 2015
© S.Sarma
14
Essence Hierarchical Architecture
IoT Symposium, Oct 2015
© S.Sarma
15
Essence Distributed Middleware APPS$2$ APPS$N$
APPS$1$ Cloud Mobile$Node$
S1$
Sn$
Sensing$&$ Sampling$
Manager$
Context$ Processing$ &$Fusion$$
Privacy$&$ se>ngs$
Communica.on$
Query$+$ Storage$
Query$ Query$&$$ Response$
Analysis$&$Processing$
AP
Communica.on$
Data$Collec.on&$ Comp.$Sampling$
Sn$
Sensing$&$ Sampling$
Manager$
Context$ Processing$ &$Fusion$
Privacy$&$ se>ngs$
Communica.on$
Query$+$ Storage$
Collabora.on$
Query$&$$ Response$
S1$
Response$
Broker$
Infrastructure$Sensing$$
…….$
Query$+$ Storage$
IoT Symposium, Oct Mobile$Node$ 2015
Manager$
© S.Sarma
S1$
S2$
…….$
Infrastructure$Sensors$
Sm$
16
Sense-‐making Using Compressed Sensing • A random sampling technique that can represent Sparse signal with few random measurements • Represents a Sparse Signal with few salient coefficients in a transformed domain • Integrates sensing, compression, processing based on new uncertainty principles
IoT Symposium, Oct 2015
© S.Sarma
17
Compressive Sensing Comparison
[Baraniuk2008] IoT Symposium, Oct 2015
© S.Sarma
18
Reconstruction##Error#(MSE)#
Collabora9ve Compressive Sensing
Legend
Sink Node(Broker)
Sampled Mobile Sensor
Mobile Node
No#of#Measurements##
Traded-‐off Number of Measurement Accuracy of Sensemaking Number of Measurement Energy Consumed in Sensing Accuracy of Sensemaking Scalability and Coverage IoT Symposium, Oct 2015
© S.Sarma
19
Execu9on Time for Sense-‐Making
IoT Symposium, Oct 2015
© S.Sarma
20
Energy Consumed in Sense-‐Making
IoT Symposium, Oct 2015
© S.Sarma
21
Applica9on: Sparse Temperature based West Nile Virus (WNV) Predic9on
[Kaggle.com]
IoT Symposium, Oct 2015
© S.Sarma
22
Research Direc9ons • Energy Efficiency – Exploit collabora9ve & compressive sensing for energy efficiency
• Incen9ve Mechanisms – Device incep9ves for par9cipa9on and collabora9on
• Privacy Regula9on – Facilitate privacy preserving incen9ves
• Heterogeneity in Mobile Cloud – Use and exploit heterogeneity of sensors and devices IoT Symposium, Oct 2015
© S.Sarma
23
Summary • Studies a new class of IoT based Sense-‐making using machine learning
– Geospa9al informa9on gathering – using combina9on of crowdsourcing and infrastructure sensing
• Proposes compressive sensing based Sense-‐making • Simula9on results are encouraging • Poten9al Extensions
– Implemen9ng a working prototype – Guide the workers to shoot photos using augmented reality – Quality assurance and cheat detec9on mechanisms
• Designed for collec9ng spa9al-‐temporal informa9on, but can be extended for event detec9on IoT Symposium, Oct 2015
© S.Sarma
24
THANK YOU
IoT Symposium, Oct 2015
© S.Sarma
25
EXTRA SLIDES
IoT Symposium, Oct 2015
© S.Sarma
26
Sensors In Mobile Phones
• MEMS & sensors for cell phones, expanding from $ 3.5 bn in 2009 to $7.9 bn in 2015 [Yole Developpement] • Smartphone sensors to be $ 6 bn business by 2016 [Juniper Research] • 44 % of the mobile phones will be smartphones in 2015 • 7x increase in mobile health apps from 2010 to 2011 • moaon sensor in smartphones and tablets will expand to $ US 2.1 billion in IoT Symposium, Oct 2015 .Sarma billion in 2011 (IHS iSuppli) 2015 w ith a 25.3 % CAGR, up from © $S1.19
27
Mobile Phone Trends • Mobile subscripaon 5.96 billion 2011 esamate • Smartphones (487.7 million) exceeding PCs (414.6 million) • More Mobile Internet Users Than Wireline Users in the U.S. by 2015 • Smartphone and bandwidth cost reduces • Smart devices contribute to more than 90% of mobile data traffic IoT Symposium, Oct 2015
© S.Sarma
28
Mobile Sensors Trends
Source: EMS Market Tracker, April 2014. IoT Symposium, Oct 2015 IHS Consumer & Mobile © M S.Sarma
29
Mobile Data Delivery Everywhere The exploding number of apps is driven by a huge upack in the number of smart devices
Smart devices contribute to more than 90% of mobile data traffic
~55%
IoT Symposium, Oct 2015
Cisco’s report 2014
© S.Sarma
30
IoT for Emergency Use Cases • Earthquakes • Hurricanes • Tornadoes • Energy/utility outages • Fire hazards • Hazardous materials releases • Terrorism/
IoT Symposium, Oct 2015
© S.Sarma
31
IoT for Emergency Response
During Fire accidents can cause electric power failure. Mobile broadcast can be used to provide direcaons to the users about rescue operaaons. IoT Symposium, Oct 2015
© S.Sarma
32
IoT for Emergency Response
Emergency situaaon Automaac Altering can be used to inform family, rescue teams, or nearby cars / passengers in case of accidents. IoT Symposium, Oct 2015
© S.Sarma
33