ict architectures for smart cities/regions - kth.se/cÅ_ictarchitectures_for... · • hololens!...
Post on 25-Apr-2018
220 Views
Preview:
TRANSCRIPT
Christer Åhlund
ICT Architectures for Smart Cities/Regions
Outline
• Introduction • A Smart City/Region ICT architecture • Sensor communication • Network and Cloud monitoring for
Communication and Processing decisions • A Smart City service – Forecasting Heat-load • Awareness of Smart City Information • Summary
A smart city in Europe
• “In Smart Cities, digital technologies translate into better public services for citizens, better use of resources and less impact on the environment.”
A smart city in Europe
• “In Smart Cities, digital technologies translate into better public services for citizens, better use of resources and less impact on the environment.”
• This is a need for rural areas as well – That is, better public services, better use of resources
and less impact on the environment.”
What does smart mean
• In this context it can be interpreted as – “simplifying” – “more efficient” – “more informed” – “controllable” – ”adaptable” – etc.
Smart from a social perspective
• A society that listens to and understands the people living there.
• Make use of the trends in engagement of citizens – Crowdsourcing – Social innovations – Research among citizens
Smart from a technical perspective
• Retrieval of information about states and activities in cities and regions to be used to manages environments in a sustainable way (ecological, social and economical)
• Requires gathering of data using sensors. Sensed data is then processed to extract information, to be informed of and manage objects
Outline
• Introduction • A Smart City/Region ICT architecture • Sensor communication • Network and Cloud monitoring for
Communication and Processing decisions • A Smart City service – Forecasting Heat-load • Awareness of Smart City Information • Summary
ICT as an enabler for smart cities and regions
User interaction
Sensorer Gateway
LTE/5G
Wi Fi
LAN Internet
Server
AAA
Outline
• Introduction • A Smart City/Region ICT architecture • Sensor communication • Network and Cloud monitoring for
Communication and Processing decisions • A Smart City service – Forecasting Heat-load • Awareness of Smart City Information • Summary
Sensing
• Requirements – Sensors need to be identified so that we can trust the
source – Sensor communication needs to be secured
• Solutions – Sensors need to authenticate – Encrypt sensor communication
Sensing
• Sensor communication solutions in this context should be able to handle: – Fixed sensors – Mobile sensors – Delay tolerant networking – Opportunistic communication
Sensing
• Authentication • Encryption
Selection of gateway
• For performance considerations we are combining information about the RSSI, round trip time, and delay.
• A policy value caclulated – PV=wRSSI*RSSIn+wRTT/RTTn+wJitter/Jittern
Some sensor authentication and communication results
• Delays experienced, time to handle authentication and battery consumption
Outline
• Introduction • A Smart City/Region ICT architecture • Sensor communication • Network and Cloud monitoring for
Communication and Processing decisions • A Smart City service – Forecasting Heat-load • Awareness of Smart City Information • Summary
Communication and processing decisions
• Where to store and process information, we need to consider: – Cloud computing performances – Communication capacity
Communication and processing decisions
• Monitoring communication and cloud platform performance
2. Network probing
Outline
• Introduction • A Smart City/Region ICT architecture • Sensor communication • Network and Cloud monitoring for
Communication and Processing decisions • A Smart City service – Forecasting Heat-load • Awareness of Smart City Information • Summary
Example of a service in a Smart City/Region
• A Bayesian Approach for Forecasting Heat Load in a District Heating System (DHS)
Forecasting Heat Load in a DHS • District Heating System(DHS)
– Heat production side – Distribution network – Heat consumption side
• DHS optimize energy production by reusing waste energy with CHP(Combined Heat and Power) plants
• CHP plants have efficiency of 65% – 90%
DHS Plant
Heated Water
Cooled Water
Residential and Commercial Buildings
CHP plant of Skellefteå Kraft
Forecasting Heat Load in a DHS
• Using a Bayesian statistical approach to develop heat load forecasting models
• Identify the influence of several parameters on the heat load forecast
• Identify the parameters with the most influence on the heat load forecast
Forecasting Heat Load in a DHS • We assume conditional independence between the parameters
influencing the heat load forecast • Naive Bayes has been shown to perform well even if there is
dependency among parameters • t = current time • h = horizon • HL(t) = Heat load • HL(t+h) = Heat load • forecast • Tout=Outdoor temp • Ts = Supply temp • Tr = Return temp
• Tdelta = Ts – Tr
• m = flow rate • Dw = day of week
• Hd = hour of day
Forecasting Heat Load in a DHS
• Two ways to learn Bayesian Network – Continuous variables – Discrete variables
• Two techniques used for
discretization: – EWD(Equal width discretization) – K-means clustering
Histogram of Heat Load of Building C during Winter Season
Forecasting Heat Load in a DHS
Equal Width Discretization • Sort continuous values from min to
max • Divide the sorted continuous range
into k intervals of equal width .
• We choose k = 5 for all parameters
K-means clustering • Divide dataset into k clusters (discrete
states) • All parameters are discretized using k =5 • Algorithm
– k data points randomly selected as centroids
– Every data point assigned to a centroid (Euclidean distance)
– Re-compute centroids for each cluster
– Repeat previous step until centroid of each cluster is fixed is fixed
K-means clusters
Proposed Model
Forecasting Heat Load in a DHS
Winter season-average accuracy of 81.23 % for all 3 buildings for HL(t+1)
Forecasting Heat Load in a DHS
Summer season-average accuracy of 76.74 % for HL(t+1)
Outline
• Introduction • A Smart City/Region ICT architecture • Sensor communication • Network and Cloud monitoring for
Communication and Processing decisions • A Smart City service – Forecasting Heat-load • Awareness of Smart City Information • Summary
How to visualise information in a Smart City/Region
• Multi-modal interaction to gain access to information
• To be made aware of situation augmented reality might be an attractive future approach
• Hololens!
Summary
• ICT solutions provide “smartness” through – Gathering of data via sensing – Adaptable communication – Data analytics for information retrieval – Information access through multi-modal interfaces
• Few deployments in place so far, often with limited to specific area
• A timely research area
Some papers • D. Granslund, C. Åhlund, P. Holmlund, “EAP-Swift: An Efficient Authentication and Key
Generation Mechanism for Resource Constrained WSNs,” To be published in Hindawi International Journal of Distributed Sensor Networks, 2015.
• D. Granlund, P. Holmlund and C. Åhlund, “Opportunistic Mobility Support for Resource Constrained Sensor Devices in Smart Cities,” Sensors 2015, Vol. 15, no. 3, pp 5112-5135.
• K. Mitra, S. Saguna, C. Åhlund, “A Mobile Cloud Computing System for Emergency Management,” IEEE Cloud Computing, vol. 1, issue 4, pp. 33-38, 2014.
• K. Mitra, S. Saguna, C. Åhlund, “A Mobility Management System for Mobile Cloud Computing,” to be printed in proceedings of the IEEE Wireless Communications and Networking Conference (WCNC) conference, 2015.
• Ss. Idowu, S.Saguna, C. Åhlund. & O. Schelén, “Forecasting Heat Load for Smart District Heating Systems: A Machine Learning Approach, ” IEEE International Conference on Smart Grid Communications (SmartGridComm 2014), pp. 554 – 559, 2014.
• S. Idowu, C. Åhlund, O. Schelén, O, “Machine learning in district heating system energy optimization,” in proceedings of IEEE International Conference on Pervasive Computing and Communications, 2014, pp. 224-227
Questions
top related