1. Sensor Network Simulation Kevin Driver, Russell Glasser, Oswin Housty

2. Class Hierarchy VibrationLightSound Environment Entity (extends thread) Node World Sensor GUI Application Display Panel Control Panel Car Monitor

3. Messaging model Nodes are implemented as independent threads Nodes are implemented as independent threads Each Node has “sendMsg”, “handleMsg” Each Node has “sendMsg”, “handleMsg” sendMsg: string → other Node object sendMsg: string → other Node object receiveMsg: Inserts the string in a queue receiveMsg: Inserts the string in a queue The thread action “ticks” every 100 MS. The thread action “ticks” every 100 MS. handleMsg called; Node dequeues one message, then performs a native action. handleMsg called; Node dequeues one message, then performs a native action.

4. Program Execution Phase 1: Click to create new sensors Phase 1: Click to create new sensors Phase 2: Sensors automatically join up in clusters; click to move car Phase 2: Sensors automatically join up in clusters; click to move car Phase 3: Visual replay of what the monitor heard Phase 3: Visual replay of what the monitor heard

5. Sensor Network Motivation Motivation Low message overhead Low message overhead Scalability Scalability Ease of implementation Ease of implementation Low Energy Adaptive Cluster Head (LEACH) Low Energy Adaptive Cluster Head (LEACH) Clusters Clusters Cluster-heads Cluster-heads Low energy Low energy Randomization Randomization

6. Sensor Network Cluster-head A Sensor 3Sensor 2 Sensor 1 Base Station Cluster-head B Sensor 6 Sensor 5Sensor 4

7. Sensor Network Our Implementation Our Implementation Cluster election Cluster election Cluster Formation Cluster Formation Cluster Communication Cluster Communication No simulation of low energy No simulation of low energy No randomization of cluster-heads No randomization of cluster-heads

8. Sensor Network Cluster-Head Election Cluster-Head Election Obtain a list of Sensor nodes Obtain a list of Sensor nodes Determine Cluster Radius Determine Cluster Radius Find Lowest ID in cluster Find Lowest ID in cluster Identify cluster-head Identify cluster-head

9. Sensor Network Cluster-head A Sensor 2 Sensor 3Sensor 1Sensor 4 Cluster Radius

10. Sensor Network Cluster formation Cluster formation Only done by cluster-head Only done by cluster-head Search list of Sensor nodes Search list of Sensor nodes Check Cluster Radius Check Cluster Radius Add to Cluster Add to Cluster

11. Sensor Network Cluster-head A Sensor 2 Sensor 3Sensor 1Sensor 4

12. Sensor Network Message passing Message passing 2 types 2 types Sensors to Cluster-head Sensors to Cluster-head Send “intruder detected” message Send “intruder detected” message Forward cluster member messages Forward cluster member messages Cluster-head to monitor Cluster-head to monitor Send “intruder detected” message Send “intruder detected” message

13. Sensor Network Cluster-head A Sensor 3Sensor 2 Sensor 1 Base Station Cluster-head B Sensor 6 Sensor 5Sensor 4

14. Monitoring the Environment Handles enqueued messages Handles enqueued messages Learns of proximity through cluster-heads Learns of proximity through cluster-heads Obtains readings from different Sensors Obtains readings from different Sensors Stores “snapshots” of anomaly Stores “snapshots” of anomaly

15. Architecture as Middleware Class Hierarchy Class Hierarchy Extensible/rich inheritance Extensible/rich inheritance Ease of development to model “real” sensors Ease of development to model “real” sensors Messaging/Protocol Abstraction Messaging/Protocol Abstraction Ignore message transmission/formatting/structure Ignore message transmission/formatting/structure Allows for different protocols Allows for different protocols Graphical tools Graphical tools Precisely, visually represent a sensor instead of updating/configuring/debugging a real one Precisely, visually represent a sensor instead of updating/configuring/debugging a real one

16. Tracking an Anomaly Monitor stores reverse distance approximation Monitor stores reverse distance approximation Snapshots stored as radiating “ripples” Snapshots stored as radiating “ripples” Snapshots show “trickle-in” of readings Snapshots show “trickle-in” of readings Approximate path through environment observed Approximate path through environment observed

17. DEMO

18. Questions ?/! ?/!