A Rescue-Assistance Navigation Method by Using the Underground Location of WSN after Disasters

Abstract

A challenging rescue task for the underground disaster is to guide survivors in getting away from the dangerous area quickly. To address the issue, an escape guidance path developing method is proposed based on anisotropic underground wireless sensor networks under the condition of sparse anchor nodes. Firstly, a hybrid channel model was constructed to reflect the relationship between distance and receiving signal strength, which incorporates the underground complex communication characteristics, including the analytical ray wave guide model, the Shadowing effect, the tunnel size, and the penetration effect of obstacles. Secondly, a trustable anchor node selection algorithm with node movement detection is proposed, which solves the problem of high-precision node location in anisotropic networks with sparse anchor nodes after the disaster. Consequently, according to the node location and the obstacles, the optimal guidance path is developed by using the modified minimum spanning tree algorithm. Finally, the simulations in the 3D scene are conducted to verify the performance of the proposed method on the localization accuracy, guidance path effectiveness, and scalability.

Keywords: anisotropic wireless sensor networks; disaster; guidance path; indoor localization; sparse anchor localization; wireless sensor networks.

Figure 1

The framework of the guidance method for underground rescue.

Figure 2

Illustration of how different features can detect the presence of the obstacles. (a–c) is the raw feature data without analysis. (d–f) is the result of processing the feature data using the hypothesis testing classifier (HTC) algorithm.

Figure 3

Missed detection probability, false alarm probability, and overall detection error probability of the HTC and SVM, showing the impact of different Sample subsets.

Figure 4

The measured and simulated power attenuation over distance in a tunnel.

Figure 5

Illustration of detecting node movement with the wall contactor: (a) is a schematic diagram of the state of the sensor node and contactor before the disaster; (b) is a schematic diagram of the state of sensor nodes and contactors when a disaster occurs.

Figure 6

System block diagram of a sensor node with accelerometer installed.

Figure 7

The example of trustable anchor judgment.

Figure 8

The average distance per hop for different densities of nodes.

Figure 9

Trustable anchor node selection scenario for single neighbor anchor node.

Figure 10

Trustable anchor node selection scenario for dual neighbor anchor nodes.

Figure 11

Trustable anchor node selection scenario for treble neighbor anchor nodes.

Figure 12

The network topology before and after disasters: (a) is the network topology before disasters; (b) is the network topology after disasters.

Figure 13

The performance evaluation of trustable anchor node selection algorithm: (a) is the visualization result of performance evaluation of the selection of trustable anchor nodes in single-layer scenarios; (b) is the visualization result of performance evaluation of the selection of trustable anchor nodes in double-layer scenarios.

Figure 14

The normalized root mean square error (NRMSE) comparison for four algorithms under different node density and different anchor ratio: (a) is NRMSE versus different node density; (b) is NRMSE versus different anchor ratio.

Figure 15

The Cumulative Distribution Function (CDF) of NRMSE under the conditions that node density is 6, and anchor ratio is 0.15.

Figure 16

Localization results and guidance paths of the method: (a) is the comparison of the actual and estimted location of the blind nodes; (b) is the visual result of the escape path search implemented by the algorithm proposed in this paper.