Data analytics, sensing and visualisation for smart cities


Krishna Kumar, kks32@cam.ac.uk
Department of Engineering, University of Cambridge




PUC-Rio, Brasil. 25 May 2017

UK National infrastructure plan

  • £466B for the next generation of infrastructure by 2020.
High quality infrastructure is essential for supporting productivity growth. Delivering the right infrastructure at a local, regional and national level, across the UK, is […] key to the government’s long-term economic plan.

Victoria station upgrade

Sensing the infrastructure

Sensing the infrastructure

Monitoring applications

Monitoring of piles

Monitoring of tunnels

Monitoring of arches

Monitoring of shaft

Wireless sensor networks

WSN evolution

Utterberry

Utterberry sensors at Eleanor Street

Data analysis from Utterberry sensors

Tunnel after dewatering

Wireless sensing & noise

WSN - Open standards

Cloud based WSN

  • Current WSN development requires complicated software set up
  • Adds to steep learning curve
  • Longer development cycle
  • Develop firmware using any browser on any device
  • Download compiled firmware
  • Deploy firmware directly to your sensor network


Hypercat

Hypercat is an open, lightweight JSON-based hypermedia catalogue format for exposing collections of uniform resource identifiers.

Hypercat - PAS212

  • PAS 212 specifies a common catalogue format that clients can use to discover data in servers that they can use.
  • PAS intentionally does not set out to solve all the challenges of data interoperability.
  • A `href` may be a URL (usually a URI)
  • Metadata 
    rel-value
    e.g., 
    {"rel":"urn:X-hypercat:rels:hasDisplacement",
                    "val":"2mm"}

Fiber optic monitoring

Distributed Brillouin Sensing

Post Office tunnel

Data analysis - tunnel movement

London Bridge Station

London Bridge Station - Overview

London Bridge Station - Piling

Sensing

  • Automatic Total Station (ATS) - absolute displacement of targets placed on the masonry structure
  • Strain sensing fibre optics (FO) - continous strain profile and crack size
  • Laser scanner - 3D understanding of structural movements

Automatic Total Station

ATS - Data analysis

It is difficult to understand the response mechanism of the structure. Small cracks are observed at quarter span – do the movements correlate with the cracks?

Laser Scan

Piecewise alignment method (PAM): finds the closest distances between points and shapes from both clouds. Uses this as an initial estimate for fitting small components of one cloud to the other cloud with minimum error. Defines movement in (X,Y,Z).

Laser Scan - Data analysis

Laser Scan - Data interpretation

Fiber Optics - Data interpretation

People movement

Smart city

Smart (monitoring complex) infrastructure

Queries with increasing level of complexity

  • At London Bridge Arch E951 area, has there been a black alarm triggered during the piles construction on dd/mm/yy hh:mm:ss?
  • Is there a crack opening greater than 5mm in the arches, when the settlement values reached 50mm?
  • What is the maximum displacement observed during tunneling (or other construction activity) within x meters from the pile.
  • Number of data readings required for a strain profile – within a tolerance of X?
  • Effect of analyzer X on scatter / noise in soil layer Y

Queries: Big, Fast and Rich

Relational database

Humidity? Vibration? 3D point cloud data? …

PostGIS

  • Geographic objects to the PostgreSQL object-relational database
  • Geometry types for Points, LineStrings, Polygons, MultiPoints, MultiLineStrings, MultipPolygons and GeometryCollections.
  • Spatial operators for determining geospatial measurements like area, distance, length and perimeter.
  • Spatial operators for determining geospatial set operations, like union, difference, symmetric difference and buffers (provided by GEOS).
v.to.db map=roads option=length type=line col=linelength units=me

Spatio-Temporal Query

  • Semantic analytics tools have primarily focused on thematic relationships, but spatial and temporal relationships are often critical components in analytical domains.
  • Current GIS and spatial database technology does not support complex thematic analytics operations.
    • GIS - modelling and analysing spatial and temporal relationships
    • Thematic entities and their relationships are not explicitly and independently represented, making analysis of these relationships difficult.

CityGML

  • Open data model and an XML-based format for the storage and exchange of data.
  • Common information model for the representation of sets of 3D urban objects

CityGML

              
                
                  1000
                  1080
                  
                    
                      model/vrml
                      
                        http://www.some-3d-library.com/3D/furnitures/Sensor434.wrl
                      
                      
                        
                          5793898.77 3603845.54 44.8

Building Information Modelling

Industry Foundation Class

              
                
                  
                    100.0
                    100.0
                    100.0

Smart cities standard

  • Smart cities standard (PAS 182) offers a handful of generic concepts (such as place, observation, metric etc.) to formulate a common language for linking data across organizations in a city.
  • The main goal is to seamlessly integrate information from multiple organisations using Linked Data (Semantic Web technology) that can be shared and edited.
  • Different from BIM, Smart Cities allow for the sharing of critical but less tangible aspects that relate to the data: such as related Assumptions and Objectives.

Linked Data - Semantic Web

Linked Data - example

Resource Description Framework

              
                

                  Utterberry measures
                  
                    Displacement.

Linked Data - Concept model

Linked Data - Time Slices

Temporal RDF

Temporal RDF extends the RDF statement from a triple to a quad where the fourth element is the valid time of the RDF statement

Graph processing

Graph optimisation

Big Data - City scale modelling

Gerry Casey, Bing Yu, Kenichi Soga, Peter Guthrie, Elisabete Silva, Krishna Kumar and Mietek Bak

SQL v NoSQL - Google directions API

London journey times

London journey times

London journey times

Traffic saturation curves

MapReduce v In-memory

MapReduce

  • Batch processing
  • Larger set of data to extract features and correlations

In-memory

  • Real-time predictions
  • GraphX: Graph parallel computing

Agent based modelling (ABM)

Graph data structure

ABM time step

Spatial v Temporal complexity

Spatial

Temporal

ABM Architecture

  • Decentralised data structures
  • Graph size: ~250k nodes and ~800k edges
  • ~1.5 GB (.gml) to 43 MB (.json.gz)

Dynamic computations

  • Shortest path at time 't'
  • Shortest path considering emissions
  • Shortest path re-computation in response to an event
  • 1 million agents in ~ 1 hour

Agent based modelling (ABM)

Agent based modelling

Cellular Automata (CA)

CA - Road degradation

CA - Road degradation

CA - Road degradation

Big Data visualisation

Tile based visualisation

WebGL - GLSL (Shading language)


GLSL - MathBox WebGL

GPU

Super computing

Alien Isolation

Thank you!!


Krishna Kumar, kks32@cam.ac.uk