1 A Delay-Aware Reliable Event Reporting Framework for Wireless Sensor-Actuator Networks Presented by Edith Ngai Supervised by Prof. Michael R. Lyu Term.

1

A Delay-Aware Reliable Event Reporting Framework for Wireless Sensor-Actuator Networks

Presented by Edith NgaiSupervised by Prof. Michael R. Lyu

Term Presentation Spring 2006

2

Outline Introduction Related Work Network Model and Objective Delay-Aware Reliable Event Reporting

Framework Grid-Based Data Aggregation Priority-Based Event Reporting Actuator Allocation

Simulation Results Conclusion

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

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

Our Focus Design of a generic framework for

reliable event reporting in WSANs Reliability in this context is closely

related to the delay, or the freshness of the events, and they should be jointly optimized

Non-uniform importance of the events can be explored in the optimization

A delay- and importance-aware reliability index for the WSANs

6

Our Framework Seamlessly integrates three key modules

to maximize the reliability index: 1. A multi-level data aggregation scheme,

which is fault-tolerant with errorprone sensors

2. A priority-based transmission protocol, which accounts for both the importance and delay requirements of the events

3. An actuator allocation algorithm, which smartly distributes the actuators to match the demands from the sensors.

7

Related Work Real-time communication protocol in WSN

SPEED [Hu et. al. 2003] real-time unicast, real-time area-multicast and

real-time area-anycast for WSN achieved by using a combination of feedback

control and non-deterministic QoS-aware geographic forwarding with a bounded hop count

8

Related Work Real-time communications in WSN

MMSPEED [Felemban et al. 2005] Multi-Path and Multi-Speed Routing Protocol for

probabilistic QoS guarantee in WSN multiple QoS levels are provided in the timeliness

domain by guaranteeing multiple packet delivery speed options

supported by probabilistic multipath forwarding in the reliability domain

9

Related Work Distributed coordination framework for WSAN

[Melodia et al. 2005] based on an event-driven clustering paradigm all sensors in the event area forward their readings to

the appropriate actuators by the data aggregation trees

provides actuator-actuator coordination to split the event area among different actuators

assumes immobile actuators that can act on a limited area defined by their action range

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, v, send data to the actuators

Anycast routing

11

Objective Reliability index

Measures the probability that that event data are aggregated and received accurately within pre-defined latency bounds

12

Grid-Based Data Aggregation

13

Priority-Based Event Reporting

We adopt a priority queue in each sensor, which plays two important roles:

1. prioritized scheduling to speed up important event data transmission

2. queue utilization as an index for route selection to meet the latency bounds

In our preemptive priority queue, the packets for the event data are placed according to its data importance and served in a first-in-first-out (FIFO) discipline

14

Delay The delay of sensor node is composed of the processing

delay, the queueing delay, the transmission delay, and the propagation delay

dtotal = dproc + dq + dtran + dprop The processing delay and the propagation delay are

typically only a few microseconds Our routing protocol allocates routes according to the data

importance Transmission delay dtran

We borrowed the idea from the SPEED protocol to estimate dtran by acknowledgement

Queueing delay dq

15

Queueing Delay The queueing delay of the highest priority

queue

16

Queueing Delay

17

Next Hop Selection

Consider node i receives new type of event data datae with

data rate It broadcasts a control message to its immediate

neighbors Every neighbors j replies with the message:

18

Next Hop Selection Node i requires that the end-to-end delay to actuator is less than the

latency bound Be

It first estimates the number of hops h from i to the closest actuator a and the maximum delay from i to j, delayi,j.

dq_max is the maximum queueing delay allowed, such that the latency bound Be can be met

19

Next Hop Selection Among the neighbors with dq_max>0, node i starts

inspecting the neighbors with λhigh=0 and λlow=0 means it is not forwarding any event data as all

next hop with λhigh>0 means it is transmitting some data with higher importance

If node i selects the next hop j with λlow>0 , then it may need to preempt some less important data

20

Next Hop Selection For each neighbor above, i calculates the maximum data

rate λi that it can forward the data to while satisfying the latency bound

The inspecting process stops when i finds enough neighbors j to forward the data, such that

21

Data transmission with Latency Constraint The latency bound Be will be updated before forwarding to

next hop

Be’ = Be – ( tdepart – tarrive) – dtran – dprop

A sensor always select a next hop that can satisfy the latency bound

If no route can meet the bound, it informs the previous hop forward the packets via another node.

In case of congestion (e.g. high priority packets flows in and preempts low priority packets), previous hop should also be informed

22

Actuator Allocation The actuators may record the

event frequency and re-arrange their standby positions periodically

Let freqg be the event frequency of the grid cell g

Estimate freqg periodically as follow:

, where freqg-1 is previous record of the event frequency in grid g

23

Actuator Allocation

24

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

25

On-Time Reachability

26

Average Delay

27

Overall Reliability

28

With Actuator Allocation

29

With Actuator Allocation

30

Conclusion We provide a distributed, self-organized, and comprehensive solution for

reliable event reporting and actuator coordination in WSAN We formulate the event reporting problem and define reliability as the

percentage of event data that can reach the destination and satisfy certain accuracy and latency constraints

We provide a distributed data aggregation mechanism, which can tolerate sensing failures and reduce network traffic

We propose a reliable priority-based event reporting algorithm with event importance. Sensors can route their data based on the affordable service rate provided by its neighbors

We further improve the efficiency of event reporting and reaction by proposing an actuator allocation algorithm. It estimates the event happening frequency in the network and balances the workload among the actuators by allocating them proper locations

Simulation results are provided to demonstrate the effectiveness of our solutions.

31

Q & A