1. Reliable Reporting of Delay-Sensitive Events in Wireless Sensor-Actuator Networks Edith C.-H. Ngai †, Yangfan Zhou †, Michael R. Lyu †, and Jiangchuan Liu ‡ † Department of Computer Science & Engineering, Chinese University of Hong Kong ‡ School of Computing Science, Simon Fraser University The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06), Vancouver, Canada, 9-12 Oct 2006.

2. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)2 Outline  Introduction  Related Work  Network Model and Objective  The Reliable Event Reporting Framework Grid-Based Data Aggregation Priority-Based Event Reporting Actuator Allocation  Simulation Results  Conclusion

3. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)3 WSAN  Collection of sensors and actuators  Sensors small and low-cost devices with limited energy, sensing, computation, and transmission capability passive devices for collecting data only and not interactive to the environments  Actuators resource-rich devices equipped with more energy, stronger computation power, longer transmission range, and usually mobile make decisions and perform appropriate actions in response to the sensor measurements

4. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)4 WSAN  Sensors and actuators collaborate sensors perform sensing and report the sensed data to the actuators actuators then carry out appropriate actions in response  Applications environmental monitoring sensing and maintenance in large industrial plants military surveillance, medical sensing, attack detection, and target tracking, etc.

5. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)5 Our Focus  Design of a generic framework for reliable event reporting in WSANs  Suggest a delay- and importance-aware reliability index for the WSANs  Non-uniform importance of the events can be explored in the optimization

6. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)6 Our Framework  Seamlessly integrates three key modules to maximize the reliability index:  A multi-level data aggregation scheme, which is fault-tolerant with error-prone sensors  A priority-based transmission protocol, which accounts for both the importance and delay requirements of the events  An actuator allocation algorithm, which smartly distributes the actuators to match the demands from the sensors.

7. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)7 Related Work  Real-time communication protocol in WSN SPEED [Hu et al. 2003]  Combines feedback control and non-deterministic QoS-aware geographic forwarding Velocity Monotonic Scheduling [Lu et al. 2002]  Packet scheduling policy that accounts for both time and distance constraints MMSPEED [Felemban et al. 2005]  Multi-path and multi-speed routing protocol for probabilistic QoS guarantee in WSN

8. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)8 Related Work  Reliable transmission with error-prone sensors Node-level fault tolerance (NLFT) [Aidemark et al. 2005]  Masks transient faults locally by using time- redundant task scheduling in nodes Bi-criteria scheduling heuristic [Assayad et al. 2004]  Uses heuristic in data-flow graph to maximize reliability and minimize runtime Routing in DTN [Jain et al. 2005]  Applies erasure code and data replication

9. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)9 Related Work  Heterogeneous sensor networks Anycast communication paradigm [Hu et al. 2004]  Constructs an anycast tree rooted at each event source and updates the tree dynamically Power-aware many-to-many routing [Cayirci et al. 2005]  Actuator broadcasts registration messages, while sensors build their own routing tables Distributed coordination framework[Melodia et al. 2005]  Sensors forward readings to the appropriate actuators by the data aggregation trees

10. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)10 Network Model  Compose of sensors and actuators  Nodes aware of their locations  Divide the network into a number of grids cell for data aggregation  A subset of nodes, referred as reporting nodes, send data to the actuators  Anycast routing

11. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)11 Objective  Reliability index Measures the probability that event data are aggregated and received accurately within pre-defined latency bounds

12. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)12 Workflow of Framework  Data aggregation  Delay- and importance-aware prioritized routing  Actuator allocation

13. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)13 Grid-Based Data Aggregation

14. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)14 Priority-Based Event Reporting  Priority queues prioritized scheduling to speed up important event data transmission queue utilization as an index for route selection to meet the latency bounds first-in-first-out (FIFO) discipline

15. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)15 Queueing Delay  The queueing delay of the highest priority queue:  The queueing delay of k th priority queue:

16. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)16 Next Hop Selection  Consider node i receives new type of event data data e  It broadcasts a control message to its immediate neighbors  Every neighbors j replies with the message:

17. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)17  The end-to-end delay to actuator should be less than the latency bound B e  Node i first estimates the advancement h i,j towards the actuator a from i to j, and then the maximum delay from i to j, delay i,j. Next Hop Selection

18. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)18 Next Hop Selection  Only neighbors with dq_max>0 will be considered as next hop  Node i starts inspecting the neighbors with λ high =0 and λ low =0 λ low =0 means it will not affect the transmission time for the existing packets in that node λ high =0 means it can be served with the highest priority  Node i calculates the maximum data rateλ i,j that it can forward while satisfying the latency bound:

19. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)19 Data Transmission  Data packets are forwarded to the neighbor with the highest h i,j and λ i,j  The intermediate nodes will update the latency bound B e before forwarding to next hop B e ’ = B e – ( t depart – t arrive ) – d tran – d prop  Sensor will update its and the routes regularly  If the latency bound is not met, another route will be selected  In the worst case, if no alternative route, sensor may inform the previous node to select another route

20. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)20 Actuator Allocation  The actuators may record the event frequency and re-arrange their standby positions periodically  Actuator allocation algorithm 1. The event frequency freq g of every grid g will be summed up 2. The field A will be equally divided in to two, denoted by A1 and A2, according to the frequency distribution 3. Each area is allocated with half of the actuators 4. The process repeats recursively for A1 and A2, until each subfield contains only one actuator

21. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)21 Actuator Allocation

22. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)22 Simulations  Simulator: NS-2  Metrics On-time Reachability Average Delay Overall Reliability  4 events 2 with high importance 2 with low importance Located in left bottom corner

23. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)23 On-Time Reachability

24. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)24 Average Delay

25. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)25 Overall Reliability

26. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)26 Actuator Allocation  Divide whole field into tree, with event occurrence probability 0.6, 0.333, and 0.067  Data rate = 60pkt/s

27. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)27 Actuator Allocation  No. of concurrent events = 10

28. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)28 Conclusion  We provide a distributed and comprehensive solution for reliable event reporting and actuator coordination in WSAN We formulate the event reporting problem and define reliability index We provide a distributed data aggregation mechanism We propose a reliable priority-based event reporting algorithm We propose an actuator allocation algorithm  Simulation results are provided to demonstrate the effectiveness of our solutions

29. The Third IEEE International Conference on Mobile Ad-hoc and Sensor Systems (MASS'06)29 Q & A