Adaptive communication for collaborative interaction in smart environments

Transcription

1 Adaptive communication for collaborative interaction in smart environments Khalil Drira, Département Réseaux et Communications (RC) Equipe s et Architectures pour Réseaux Avancés (SARA), Toulouse, France RT Days, June 21, /07/ Towards open & cognitive IoT Buy & track shipped plants Track Things in real time Communication Protocols interoperability Plants act on the irrigation system to water themselves Let Things talk & understand each others (M2M) Integration of heterogeneous Systems & Semantic Reasoning Data interoperability Provide on line content View images of plants & gardens 09/07/ Allow the remote control of Things Persons remotely monitor and control garden irrigation facilities Let Things become smart & cooperative Cooperation between systems : irrigation, weather & botanic to adapt watering 2 1

2 LAAS/SARA R&D directions for IoT challenges 1 Vertical Fragmentation Power Management Increasing Complexity Security Network Misalignment Common s & Horizontal Architectures Semantic Interoperability: Protocol level Data level Energy Saving: Protocollevel, Middleware level LPWAN (LoRA, NB-IoT) Autonomic Management: Semanticenabled Dynamic deployment discovery & integration Detect anomalies Softwarized & & Prevent DoS Virtualized attacks: Unsupervised Machine / Reinforcement / Deep Learning Networks: SDN, NFV, Overlays 1 The internet of things: Key applications and protocols O.Hersent, D.Boswarthick, O.Elloumi John Wiley & Sons, /07/ Enabling cross-domain interoperability Building Energy Security Building Energy Security Vertical Fragmentation Existing ICT infra Existing ICT infra Exis9ng ICT infra Invert the pipe Common service platform Converged ICT infra Dedicated Dedicated Dedicated M2M M2M M2M > Horizontal Interoperability : International standards onem2m, ETSI SmartM2M. > Other initiatives & projects: Allseen, EU projects,etc. 09/07/

3 Semantic gap breaks IoT horizontality rqrzflvb7si/5kuxvqodlnmxuu7 TEMP-AZ1299B, 17, C, T <observation> <ID>TEMP-AZ1299B</ID> <data>17</data> <unitofmeasure>c</ unitofmeasure> <timestamp>126738</timestamp> </observation> <measurement> <device>temp-az1299b</device> <value>17</value> <unit>c</unit> <time> t192030</time> </measurement> observation Tag- level1 measurement Data Hexa/Binary Text Keywords Knowledge Description Taxonomy Ontology 09/07/ /07/ Semantic IoT: The goals > Effective data interoperability between and applications without any prior agreement. > Generic interworking and automated management of. > Semantic discovery and data querying. > Semantic matching and binding of and apps. > Semantic reasoning to infer new knowledge from facts. > Better monitoring and understanding of the surrounding environment. > Smart decisions to dynamically adapt to environment changes. 6 3

4 LAAS s OM2M: IoT sw service pla3orm > T h e re fe re n c e s ta n d a rd -c o m p lia n t o p e n s o u rc e p la tfo rm : O M 2 M.o rg > O n g o in g e x te n s io n s (+ 1 0 K lin e s ) b y O ra n g e & D e u ts c h e Te le k o m > E x p e rim e n ta l d e p lo y m e n t a t L A A S : A D R E A M /S m a rt A p p a rt/r o b o ts (w ith D O /R IS ) > D e p lo y e d to p ro te c t th e g a te w a y o f th e A D R E A M b u ild in g fro m d a m a g e c a u s e d tw ic e b y o v e rs ize d tru c k s L A A S -C N0R9S/ 0 7 / IoT-O: LAAS ontology for IoT > Validated and disseminated in Communication Standards SI, IEEE Comm. Mag. Dec 2015 > Enhanced, implemented, experimented since 2016 (W ile y IT L ) > Reused for contribution to the international standard onem2m Base Ontology (/Alcatel-NOKIA) : OneM2M base ontology proposal MAS R01, June 2015 L A A S -C N0R9S/ 0 7 /

5 Highlights, Deployments, Experiments, Hackathons LAAS Eclipse OM2M V1, V2 Startup hosted by IoT Valley Toulouse L A A S -C N0R9S/ 0 7 / Semantic models: Recent Publications An Autonomic Cognitive Pattern for Smart IoT-based System Manageability: Application to Comorbidity Management, ACM TOIT to appear in A model-driven methodology for the design of autonomic and cognitive IoT-based systems: application to healthcare. IEEE Transactions on Emerging Topics in Computational Intelligence, Vol.1, N 3, Jun Cache freshness in na for the internet of thin Journal, Janv. 2018, Towards semantic data interoperability in onem2m standard. Communication Standards SI. IEEE Communications Magazine, Dec Contribution to OneM2M Standard, Mar OneM2M base ontology proposal. L A A S -C N0R9S/ 0 7 /

6 Machine Learning: Recent Publications Unsupervised Network Intrusion Detection Systems: Detecting the Unknown without Knowledge, IEEE Network, 26 (1), IoT Attack Detection with Deep Learning, ISCIS Security Workshop Big Data for Autonomic Intercontinental Overlays. IEEE JSAC, 34(3), Scalable, Self-healing and Selfoptimizing Routing Overlays, IFIP Networking /07/ Implementation scenario 09/07/

7 Thanks 09/07/ AI Software-Defined Networks Network Virtualisation Open Interoperable IoT Architectures Related activities of SARA: executive summary SYNERGY Transversal Axis Recent Projects Platforms CNES Orange Panacea TOUIX RTRA/ CYPHYS A2NETS e-horizon AATAC S2C2 LAAS/OPA IDEX/Chair ROSACE DGA Endeavor IMAGINE E-health (with LIST/LU) Fleet Management (with ATOS) Smart Metering (with VTT/FI) Robotics (with RO/RIS) Dynamic Manufacturing Networks (with Airbus group) Rescue Systems (with RO/RIS, IRIT, ONERA) Tools & FW Reinforcement Learning Unsupervised learning Game Theory SemanLc & Ontologies FM, Graphs & G. Grammars Traffic Management & Optimization Scalability Management GMTE Theories &Techniques 09/07/2018 FACUS 14 OM2M V0.8 OM2M V1.0 FrameSelf 14 7

8 Semantic IoT vs Semantic Web > Semantic IoT has more requirements and constraints than Semantic Web. > It requires continuous: monitoring, pre-processing, filtering, aggregation, annotation, and integration. 09/07/ Mastering complexity by semantic reasoning Autonomic Manager [Kephart 03] Monitor Analyze Knowledge Plan Managed Element Execute Challenges for Autonomic Mgt in IoT: Generic solutions for autonomic management of IoT systems. Ontology for semantic reasoning: selfconfiguration of Increasing Complexity 09/07/

9 Self-configuring IoT (1/4) 1. Monitoring Runtime discovery of and update of the IoT-O ontology instance. M A K P E Server SERVER_NODE Gateway 1 WAN Gateway 2 GATEWAY1_NODE Sensor TEMPERATURE_SENSOR STATE_SERVICE GATEWAY2_NODE HEATER_ACTUATOR CONTROL_SERVICE Temperature sensor Heater 09/07/ TEMPERATURE QuantityKind 17 Self-configuring IoT (2/4) 2. Analyzing Apply seman5c rules to find relevant matching between and actuators. M A K P E Server SERVER_NODE Gateway 1 WAN Gateway 2 GATEWAY1_NODE Sensor TEMPERATURE_SENSOR STATE_SERVICE GATEWAY2_NODE HEATER_ACTUATOR CONTROL_SERVICE Temperature sensor Heater 09/07/ TEMPERATURE QuantityKind 18 9

10 Self-configuring IoT (3/4) 3. Planning Query the ontology instance to find service operanons of matched to create acnons. M A K P E Server SERVER_NODE Gateway 1 WAN Gateway 2 GATEWAY1_NODE Sensor TEMPERATURE_SENSOR STATE_SERVICE GATEWAY2_NODE HEATER_ACTUATOR CONTROL_SERVICE Temperature sensor Heater 09/07/ TEMPERATURE QuantityKind 19 Self-configuring IoT (4/4) 4. Executing Convert actions to HTTP requests and create required device subscriptions on the platform. M A K P E Server SERVER_NODE Gateway 1 WAN Gateway 2 GATEWAY1_NODE Sensor TEMPERATURE_SENSOR STATE_SERVICE GATEWAY2_NODE HEATER_ACTUATOR CONTROL_SERVICE Temperature sensor Heater 09/07/ TEMPERATURE Quan:tyKind 20 10

11 Ambient Intelligence: the scope of challenges 09/07/