1. Fast Localization for Emergency Monitoring and Rescue in Disaster Scenarios Based on WSN
SPEAKER：Jyun-Ying Yu
ADVISOR：DR. Kai-Wei Ke
DATE：2018/05/04

2. Outline INTRODUCTION RELATED WORK MATERIALS METHODS
SIMULATION and ANALYSIS
CONCLUSION

3. INTRODUCTION
When disaster strikes, it is difficult to rescue because of environmental impact.
It is more appropriate and economic to deploy WSN than wired equipment, especially in dangerous environments.
Base on RSSI to obtain distances, proposing an improved weighted centroid algorithm to get a more accurate position without further cost and energy consumption.

4. Related work

5. Angle of Arrival (AOA)

6. Time of Arrival (TOA)

7. Received Signal Strength Indicator (RSSI)
On Free Space propagation model ●

8. MATERIALS

9. MATERIALS
The proposed emergency monitoring and rescue system is consisted of three layers.
Sensor node
Anchor node
Base station

10. Wireless nodes Microprocessor：Atmega128 RF transceiver：Chip CC2420
Software system：TinyOS

11. Anchor nodes GPS Camera Wireless node Storage battery
Solar charging panel

12. METHODS

13. An improved localization algorithm based on RSSI
The classical radio propagation model considering pass lost and shadow effect is presented as follows: ●

14. (Cont’d)

15. (Cont’d)

16. (Cont’d)
The distance dˆi between sensor node and anchor node can be computed：

17. An improved Weighted Centroid Localization (WCL) with adaptive correction
Viewing a sensor network with n sensor nodes and m anchor nodes in the disaster ruins.
The positions of the sensor nodes are unknown, whereas the positions of the anchor nodes are known.
In the first phase, all anchor nodes broadcast their position (xj, yj) to all sensor nodes within their transmission range.
Upon reception of k (k ≤ m) anchor positions, the coordinate (xˆi, yˆi) of the sensor node is estimated according to the following formulas.

18. (Cont’d)
The weight wi j is a function depending on the distance and the characteristics of the sensor nodes’ receivers. The nodes with shorter distances are more weighted than the ones with longer distances.

19. (Cont’d)
In other words, the weight and the distance are inversely proportional. The weight is calculated by the following formula:

21. To solve the problem
Proposing a new method named L-WCL to determine the weight given in：
However, the accuracy of RSSI-based distance measure tends to be affected by the environment in practice. If we calculate the distances between nodes only using RSSI without correction methods, it may lead to a high localization error. 21

22. (Cont’d)
A calibration factor λi j is defined to correct the estimated distance using the following formula (p.22), where dˆ i j is the calculated distance between sensor node Si and anchor node Aj by the formula (p.16), di j is the corrected distance .

23. L-WCL-AC

24. SIMULATION and ANALYSIS
Parameters are set as follows.
Deploying 10 anchor nodes and 90 sensor nodes in a 100m*100m mission space.
The anchor nodes are uniformly distributed in advance with known positions, while the sensor nodes are tossed at random.

25. (Cont’d)
The path loss model parameters PT x, K and η have been estimated from the collected RSSI measurements according to a mean square error criterion in the actual experiments.
PT x +K = −45.14
d0 = 1m
η = −1.37
σ2 = 5
The mean localization error of sensor nodes：

30. CONCLUSION
The paper has presented a new localization algorithm which is suitable for emergency monitoring and rescue in disaster scenarios based on WSN. The algorithm is mainly divided into three stages.
Compared to WCL, L-WCL-AC has higher accuracy and is more suitable for localization through the sensor network with sparse anchors. In the future work, the system will be deployed in several separated disaster areas to test its robustness and efficiency.

31. Reference
M. Lu, X. Zhao, and Y. Huang, “Fast localization for emergency monitoring and rescue in disaster scenarios based on WSN,” th International Conference on Control, Automation, Robotics and Vision (ICARCV), 2016.
Y.-L. Chu, J.-R. Tzeng, Y.-P. Cheng, and M.-T. Sun, “Density-Adaptive Range-Free Localization in Large-Scale Sensor Networks,” st International Conference on Parallel Processing Workshops, 2012.
P. Levis, “TinyOS: An Open Operating System for Wireless Sensor Networks (Invited Seminar),” 7th International Conference on Mobile Data Management (MDM06), 2006.
A. Goldsmith, “Wireless communications,” Cambridge university Press,

32. Thanks for Listening