SensXplore - Semantic Web Challenge
Contents
1. etc are captured under SFR Fig 2b This inherits from action concept of OntoSensor and provides a good platform for API service exposure for meters SDR Fig 2c represents data and basic parameters as measured by meters like power voltage current etc This tree starts frm OntoSensor concept called Parameter and Quantitiy SDER Fig 2d starts from OntoSensor concept of CapabilitiesDescription where all types of communication protocol relevant to meters are listed MeasuringDevice Y Thing D action Sensor i 0 i oe Rotation ction r Electrical Y B SensorAction OElectroMechanical Y D EnergyMeterFunction T P Induction D ReadFunction 2 il EnergyMeter 2 2 BasicReadFunction cM Pie steamaanetie amp OInferredReadFunction g D WriteFunction os SolidState B BasicWriteFunction iD Humidity B InferredWriteFunction a SHR b SFR es Cooricient Ww CapabilitiesODescriptiaon E sleet ae H WS PlatformCapabilities koii a on allege Fr crud See ae es te te 5 0 Y LongHaul Jia Oe o GPRS Lae oo Coe i i i IPBased i LS Sensumetion i tate a MIB us oe ame ie ae Cer i eee gt Modbus he ee cCceleration Parane ter i a Optical pase pl a f Serial a Distance H D WirelessHuURT Sg as pea E ee ftyP ara her 3 aia WirelessMBus oo a TAE H wrirelessmodBus o D FigBee2. 6 which is our proprietary platform for sensor based intelligent systems This tool is capable to cater needs of sensor manufacturers in this case meter manufacturer sensor based service providers here utility service providers who install meters at households as well as query makers who in this case are utility service providers consumers and data researchers There are three modules in the tool Purpose and uses of these modules is explained below A details description of how to use the tool with proper screenshots is given in Appendix 6 Sounak Dey Ranjan Dasgupta Arpan Pal and Prateep Misra Sensor Registration This module is meant for sensor manufacturers As on tology already holds details knowledge base of a type of sensor its hierarchy general factsheets of that type of sensor other details like capabilities working range accuracy data communication protocols and interdependencies of these parameters manufacturers will only have to select the right position from sensor hierarchy this is where their sensor fits and select the features capabilities or semantic categories relevant to their sensor and will register the new sensor in the CPS In this way interoperability in terms of sensor definition would be there Also higher level semantics can be embedded with metadata level for future pro cessing In future this work can be extended to accommodate modification in sensor feature set or hierarchy in ontology against m
3. i ake Single T re one D are T aaa E i k ElectricalmMetercapability c SDR d SDER Fig 2 Different modules of ontology SensXplore A sensor discovery tool 5 Meter specific concepts are logically parked into DSR placeholder This can be spread across all above four logical divisions Concepts like MeterConstant UserConsumption three phase meter falls under this category This model can be followed for any other type of sensor and that sensor can be semantically described using this model 4 Tool Design and Description Based on the requirements as specified in section 2 2 and taking sensor ontology described in section 3 as a backend platform a tool sensXplore has been designed and implemented This is our second contribution Sensor example for both ontology and tool is electric meter Architechture diagram for the tool is given in Fig 3 2 USER SENSOR CONCEPT Query DEVELOPER Sensor service discovery capability exploration Concept registration oe Ay ontology Query processor extension Jf GC fo E fo Brrr rrr rrr rrr reersere Register new Sensor sensor i i Sensor i Registration time series Sensor i cau Metadata a o M Bus SENSOR MANUFACTURER Sensors Selection of sensor based A on metadata ex l SERVICE wil PROVIDER Fig 3 Architechture of sensXplore The tool is hosted in RIPSAC Real Time Integrated Platform Analytics and Services for City
4. SensXplore A Tool for Sensor Discovery using Semantics with Focus on Smart Metering Sounak Dey Ranjan Dasgupta Arpan Pal and Prateep Misra Innovation Labs Kolkata Tata Consultancy Services Ltd India sounak d ranjan dasgupta arpan pal prateep misra tcs com http www tcs com Abstract Sensor capability discovery services discovery and selection of right sensor as per application requirement is very important aspect for sensor systems Structured semantic representation of the same in an ontology would be helpful for manufacturers application developers and sensor deployment users This paper describes an online tool called sen sXplore which uses such ontology as a backbone to serve the aforesaid purpose Keywords Sensor Interoperability Discovery Tool Ontology 1 Introduction Present day cyber physical systems CPS are becoming more and more depen dent on sensors With frequent use of sensors there come challenges to be solved with respect to sensor data semantics sensor capabilities exploration tasking discovery etc Based on user classification and technology SSN XG Report 1 has clearly divided the challenging use case sets in four broad categories 1 Data Discovery and Linking 2 Device discovery and Selection 3 Provenance and Di agnostics and 4 Device Tasking and Operations As the names suggests first category is more focused on sensor data semantics and relation discovery while the second one i
5. To be appeared IEEE 2013 4 Niles I Pease A Towards a standard upper ontology In Proceedings of the international conference on Formal Ontology in Information Systems Volume 2001 ACM 2001 2 9 5 SWE OGC Sensorml standard http www opengeospatial org standards sensorml 2013 Online accessed 20 Sept 2013 6 MISRA P PAL A PURUSHOTHAMAN B BHAUMIK C SWAMY D SUB RAHMANIAN V S KAR D NASKAR S ADAK S GHOSH S et al A com puting platform for development and deployment of sensor data based applications and services May 24 2013 WO Patent 2 013 072 925 Appendix Online Link of sensX plore sensXplore tool is publicly available in the link http ripsac2 web2labs net ontologyUi User Manual of sens Xplore Step 1 On going to this link there will be a prompt for username and pass word User will have to enter username tcs and password tcs for further progress into the demo of the tool Step 2 on successful log in demo user will see a homepage as shown in Fig 4a There are three options to choose from namely Sensor Registration Sensor Provisioning and Sensor Discovery Step 3 On clicking Sensor Registration link screen as shown in Fig 4b will appear This screen is for use of manufacturers They will search the tree in the left pane to find out desired sensor type from the hierarchy of sensors On selecting one type of sensor text boxes on t
6. anufacturer s suggestions subject to a proper review by domain experts Sensor Provisioning When sensors are registered and are ready to be de ployed by CPS enablers then they can select correct sensor by searching its feature performance capabilities etc and deploy and activate the sensors in actual fields To elaborate with an example when multiple electric meters are registered in the system by multiple makers electric supply company can search for best meter that fits their requirement and can deploy and provision it to households by attaching customer details with the meter id and defining a space for data storage from that meter All these can be done without knowing much intricacies of standard format etc followed by manufacturers Sensor Discovery Once the sensor is registered and provisioned in the system then the obvious requirement will be gathering time series data and using them to analyze and answer complex queries Another angle of requirement will be to query the sensor itself about its factsheet metadata capabilities etc to help establish the reliability of data This module focuses on second angle and just puts a pointer to data for the former part Queries here may be meter centric like retrieve all the AMIEnabled meters whose firmwire can be upgraded or customer centric like find all customers having Pbased meters In future this module will look into handling more complex semantic q
7. comprehensive knowledge base or interface which can be used for sensors or at least for meters There are some ontologies like SWAMO CSIRO SDO OntoSensor etc Of them CSIRO lacks a clear hierarchy of sensors SWAMO lacks clear category separation and top classes are a bit mixed up SDO focuses only on data though and have very limited resource to work with Of all these sensor ontologies OntoSensor found to be a good starting platform for this requirement as it inherits Suggested Up per Merged Ontology SUMO 4 by IEEE and also ISO 19115 standard in some its relations For e g Sensor Platform and Event concepts of OntoSensor extend the Measurement Device Transportation Device and Process con cept of IEEE SUMO respectively Also it incorporates some concepts of SWE OGC standard for sensor description language SensorML 5 2 2 Tool Requirement In absence of what proposed in section 2 1 stakeholders of CPS face many practical problems as discussed here Manufacturers while releasing a new sensor of some category have to de scribe its metadata fact sheet capabilities etc in their own way in absence of a standard description tool Having a standard tool backed by a stable ontology which can even be modified by experts as and when required on manufacturer s request would be a nice tool In absence of standard tool when user of such sensors want to deploy them into practical fields then they
8. epresentation technique and using it as a semantic tool to enrich both sensor data and metadata Though there are many top level approach for describing sensors it has been observed that OntoSensor 2 goes close to cover all the aspect of a sensor Taking OntoSen sor as a starting point the ontology has been extended for energy sensor and created one example ontology which carries modularized approach to describe sensor hierarchy sensor data representation sensor functionalities and some sen sor domain specific representations All these have been discussed in our earlier work 3 and have been implemented for energy sensors Based on this ontology a tool named sensXplore has been made which have an interface for manufacturers to enter details of their sensors and an interface for sensor deployment users who can select and provision the sensor at per requirement and finally one query interface which will enable users and application developers to search sensor capabilities services data semantics etc 2 Related Works 2 1 Ontology Selection To capture metadata fact sheets capabilities etc of sensors there needs to be a semantic knowledge base which will hold all these details in a structured way SensXplore A sensor discovery tool 3 in an ontology so that sensor deployment team can search a sensor easily and efficiently based on given constraints and requirements Prior art search says that there is no such ready
9. he right hand side will be enabled and user have to enter sensor model name maker name load capacity and have to select what capabilities the sensor has from the list of checkboxes When done with data entry and selection user will press Register to register the sensor 8 Sounak Dey Ranjan Dasgupta Arpan Pal and Prateep Misra Step 4 Now going back to homepage in Fig 4a user will select Sensor Provisioning link This screen is as shown in Fig 4c It is for use of sensor deployment users User will select one of the registered sensors from the left pane Upon selecting the sensor the capabilities and type of that sensor will be populated in right portion of the screen and user have to enter customer name address id location in latitude longitude format eg 88 25 22 78 and sensor id Upon data entry user will press Submit and the sensor will be provisioned for that customer id Step 5 Again on going back to homepage Fig 4a user may select Sensor Discovery This screen as shown in Fig 4d is for sensor service and capabilities explorers User will use checkboxes to choose from sensor capabilities and query can be done based on customer id or sensor id On clicking Submit output will be shown Also queries like which customer has which meter can be done from this screen Note Pressing the Cancel buttons in all screens will stop whatever op eration was being done in that
10. le of smart meters henceforth energy sensor it has been seen that there are many standards like DLMS COSEM ANSIC12 etc which meter manufacturers generally follow but there is huge variation in terms of capabil ity of such meters for e g measurement parameters while some meter measures values of active energy reactive energy some others gives energy only measure ment precession some meter measure values up to multiple decimal points com munication protocol followed there are different protocol like Modbus MBus serial port optical port zigbee etc data storage and analytics capacity some has data storage chips built in and so on Before deployment of a meter one has to go through all these details and have to decide the best meter that fits the re quirement Also manufacturers do not have a common platform to describe their meters so that it can be used by those who will deploy meters in households Thus it is obvious that there is a requirement for a common knowledge base which can hold details of sensors as provided by manufacturers its spec ification accuracy factsheet performance under constraints etc Again that knowledge base should be semantically searchable by sensor deployment users Also it should help explore sensor services Application developers should also use the same knowledge repository Looking at the nature of the problem an obvious way to address the problem is using ontology as a data and knowledge r
11. need to check out ev ery individual sensor s metadata sheets separately to understand its capabilities precession etc This can again be reduced by a tool that will help selecting sensors by categories types capabilities way of functioning communication protocols etc Last section consists of developers who use services of a CPS and query sensor data and metadata analyze them and produce meaningful inference or actions A meaningful semantic query tool will help their need to much extent 3 Proposed Ontology Design and Implementation Energy sensor has been selected and used as an example for which OntoSensor has been extended into much more details for some classes defined there Also this richer ontology is divided into some logical modules as shown in Fig 1 like sensor hierarchy SHR sensor functionalities SFR sensor data SDR and sensor data exchange SDER There is also a placeholder DSR for keeping domain specific information Meters can be classified and divided into hierarchy based on any of 1 mea surand 2 sensor material 3 application 4 cost 5 accuracy 6 output signal etc In our case measurand has been taken as the key feature based on which meter has been categorized in SHR Fig 2a 4 Sounak Dey Ranjan Dasgupta Arpan Pal and Prateep Misra a Fig 1 Logical modules in ontology All functional capabilities and services provided by meters like Reading Data SendData
12. particular screen and will take back users to home page Fig 4a Pressing Home button will take user to same home page Pressing Log Out from any screen will log out user from the system aM Measuring De 4 Electrical li Electro Meche 4 D Induction D Electro Magne Solid State Sensor Reaistration Sensor Provisioning Sensor Discovery You can Register a New Sensor here You can Provision a New Sensor Here You can Search for a Sensor here lt EEEREEEEEHEHEHEE a Home page b Sensor registration Registered Sensor List c Sensor provisioning d Sensor discovery Fig 4 sensXplore Screens
13. s related to sensor services and capabilities exploration and se lection of sensor as per application requirement Third category is an extended subtle set of second category and last one relates to low level commands and operation Though data semantics and discovery is a very important topic with respect to CPS but sensor device and capability discovery category 2 is more impor tant for those who deploy sensors in a system Exploring capabilities and fitness of a sensor with respect to a particular application is important as this will decide the system performance and reliability Moreover quality and usefulness of data way of tasking and operation of system scalability of system etc are dependent on the kind of choice of sensor is made before deploying So those who are responsible for selecting and deploying sensors for a particu lar use case need to understand and search different capabilities and factsheets of 2 Sounak Dey Ranjan Dasgupta Arpan Pal and Prateep Misra sensors and evaluate performance of sensor under given constraints from differ ent manufacturers following specification standards defined by various standard bodies All these are highly variable over time so one deployment cannot entirely help next deployment So a reinvention process is waiting every time a new de ployment comes If the deployment goes wrong then that also creates problem for application developers who use data from sensors Taking examp
14. ueries For example semantic queries like Find high consumption customers in a particular area during peak load time in Summer contains different semantic terms like High Consumption Peak Load Summer etc These terms are defined either in the base ontology or some other ontology that have been imported and reused This knowledge base and interrelations and defined set of rules in the system will finally help infer the final output of the query 5 Future Scope More structured approach for ontology creation can be done usign SSN philoso phy which revolves around concepts like stimuli feature of interest observation etc as explained in 1 Also a Query Processing unit would be introduced which SensXplore A sensor discovery tool 7 will help break down categories of queries translate natural language queries into standard sparql queries make a query plan etc References 1 SSN XG Ssn xg final report http www w3 org 2005 Incubator ssn XGR ssn 20110628 2011 Online accessed 20 Sept 2013 2 Russomanno D J Kothari C Thomas O Sensor ontologies from shallow to deep models In System Theory 2005 SSST 05 Proceedings of the Thirty Seventh Southeastern Symposium on IEEE 2005 107 112 3 Dasgupta R Dey S A comprehensive sensor taxonomy and semantic knowledge representation Energy meter use case In Proceedings of the 7th International Conference on Sensing Technology
Download Pdf Manuals
Related Search