Integrated modeling and simulation framework for wireless sensor networks


Published on

Presentation at the 3rd IEEE Track on Collaborative Modeling & Simulation - CoMetS'12

Published in: Education
  • Be the first to comment

  • Be the first to like this

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Integrated modeling and simulation framework for wireless sensor networks

  1. 1. Integrated Modeling and Simulation Framework for Wireless Sensor Networks Baobing Wang and John S. Baras Institute for Systems Research Department of Electrical and Computer Engineering University of Maryland, College Park, USA WETICE/COMETS, Toulouse, France June 27, 2012
  2. 2. Content• Motivation• Contributions• System Framework and Design Flow• Model Libraries• Case Study• Conclusion 2
  3. 3. Motivation• WSNs as cyber-physical systems – Physical environments, physical platforms – Communication protocols – Computation algorithms• Drawbacks of current design methodology – Ad hoc design approaches – Test only limited design alternatives – Lack of reusability – Little consideration of the interaction between continuous-time domain and event-triggered domain 3
  4. 4. Motivation• Support system design for heterogeneous WSNs – Hybrid design framework required: event- triggered and continuous-time dynamics – Component reusability – Multiple optimization objectives: trade-off analysis and design space exploration – Mission-critical applications: model-checking 4
  5. 5. Contributions• Propose a model-based system design framework for WSNs – Integrate both event-triggered and continuous-time dynamics – Provide a hierarchy of system model libraries• Propose a system design flow within our model- based framework – Based on an industry standard tool – Simulation codes (Simulink and C++) are generated automatically – Support trade-off analysis and optimization 5
  6. 6. System Framework• Model libraries – Application Model Library – Service Model Library – Network Model Library – Physical System Model Library – Environment Model Library• Development Principles – Event-triggered: Statecharts in SysML – Continuous-time: Simulink or Modelica 6
  7. 7. System Framework Distributed Computing Communication and Sensor Database Physical World 7
  8. 8. System Design Flow 8 8
  9. 9. System Design Flow• Trade-off analysis and design space exploration – Each component is described with performance index – Parametric Diagrams – Parametric Constraint Evaluator or CPLEX• Simulate in Matlab/Simulink – All SysML blocks are transformed into a single S-function – Generate Simulink source file• Simulate in IBM Rational Rhapsody – Generate C/C++ source code – Statechart animation – Interactive simulation 9
  10. 10. Model Libraries• Physical Platforms – Battery, CPU, sensors and transceiver 10
  11. 11. Model Libraries Example: behavior model of a transceiver using Statechart in IBM Rhapsody 11
  12. 12. Model Libraries• MAC Layer Components – Low Power Listener: adjust radio’s power state based on channel activity – CSMA/CA Channel Access: gain channel access right in CSMA/CA mechanisms – CSMA/CA Sender: send one packet with retransmissions in CSMA/CA mechanisms – MAC Controller: specify the control logic of a MAC protocol (ports are defined, but behavior should be customized for each protocol) – Slot Manager, Queue Manager, TDMA Sender, Receiver … 12
  13. 13. Model LibrariesExample: IEEE 802.15.4 unslotted mode for Controller 13
  14. 14. Model Libraries• Physical Environment – Modeled using Simulink or Modelica – Built using the Embedded Coder to generate C/C++ codes – Imported as SysML blocks – New environment information are pushed to event- triggered blocks periodically• Wireless Channels – Radio propagation models, channel fading models and BER under different modulation schemes – Currently support: free space model, UDG model, ITU indoor model and Rayleigh fading model 14
  15. 15. Case StudyData Center Living Room 15 15
  16. 16. Case Study• Simulation scenarios – Wireless + No Pipe: disable pipe, send data wirelessly, measure period is 5 seconds – Wireless + Pipe (5s): similar to above, but pipe is enabled – Wireless + Pipe (60s): similar to above, but measure period is 60 seconds – Wired + Pipe: temperature data are available immediately and directly 16
  17. 17. Case Study 17
  18. 18. Simulate Results 18
  19. 19. Conclusions• We proposed a model-based system design framework for WSNs• The proposed system design flow can integrate both continuous-time and event- triggered modules, and study the system performance using generated codes• Composability and reusability are demonstrated through a case study 19
  20. 20. Thank You! 301-405-6606 Questions? 20