KR100571124B1 - Slope disaster monitoring system - Google Patents

Abstract

The present invention relates to a slope disaster monitoring system, and more particularly, to a slope disaster monitoring system for monitoring a disaster by detecting data input of an abnormal state using an auto-associative neural network.

Disaster detection system according to the present invention is a disaster detection device for receiving the disaster detection data received from a plurality of detection sensors and transmits to a monitoring server, the memory and the device for storing the detection data output from the detection sensor and the A switching unit for selectively outputting one of the sensing data read from the memory and the sensing data input from the sensing sensor, and a data transmitter for wirelessly transmitting the data output by the switching unit to the monitoring server and the sensor. Outputting a control signal of the switching unit to read and output the sensing data stored in the memory every predetermined period when the sensing data is less than the prescribed value, or transmitting the sensing data in real time when the sensing data is more than the prescribed value. It characterized by including the above signal detecting portion for outputting a switching control signal group.

Accordingly, the slope disaster monitoring system according to the present invention determines whether there is an abnormality of sensing data inputted by a plurality of disaster detection devices, and transmits the corresponding data to the monitoring server in real time when data of an abnormal pattern is input, thereby early detecting the occurrence of a disaster. It has the effect of quickly suppressing.

Neural network, slope disaster, sensor, web

Description

Slope disaster monitoring system {Calamity monitoring system}

1 is a general schematic diagram of a slope disaster monitoring system according to an embodiment of the present invention.

2 is a schematic block diagram of a disaster detection apparatus according to an embodiment of the present invention.

3 is a schematic block diagram of a layer structure of an automatic computation neural network of a disaster detection apparatus according to an embodiment of the present invention.

4 is a schematic block diagram of a monitoring server according to an embodiment of the present invention.

5 is a schematic block diagram of a monitoring server according to another embodiment of the present invention.

6 is a schematic block diagram of an automatic computational neural network of a nonlinear technique in accordance with an embodiment of the present invention.

<Explanation of symbols for the main parts of the drawings>

100. Disaster Detection Device 110. Detection Sensor

120. Autonomic Neural Networks 121. Input Layers

122. Mapping Layer 123. Bartlek Layer

124. Demapping Layer 125. Output Layer

130. Memory 140. Switching unit

150. Data transmission unit 160. Error signal detection unit

200. Monitoring Server 210. Data Receiver

220. Disaster signal detector 230. Decryptor

240. Autonomic Neural Networks 241. Bartlek Layer

242. Demapping Layer 243. Output Layer

250. Disaster Information Database 260. Video Output

270. Voice output unit 280. Web service provider

300. User terminal

The present invention relates to a slope disaster monitoring system, and more particularly, to a slope disaster monitoring system for detecting an abnormal data input by using an auto-associative neural network and monitoring it when a disaster occurs.

In general, a remote disaster monitoring system installs CCTV cameras or detection sensors in a large number of areas where a disaster is likely to occur and detects a disaster early and quickly suppresses the disaster. Is a system to notify.

On the other hand, such a disaster prevention system is generally installed in a place that is not easily accessible to the monitor. Therefore, it is almost impossible for the observer to stay at the measurement site and collect the observation data in resources and time, so most disaster prevention systems receive the observation data input from the site's monitoring device remotely or remotely by wire or wirelessly. Form. Many sensors are installed in the field to increase the reliability of the measurement data. Measurement data input from the sensor is transmitted to the server at predetermined intervals for efficient disaster prediction.

Domestic and international programs related to the slope disaster monitoring system include Slope Stability Analysis Software (STABL) developed by Purdue University and Slope Stability Software developed by ASCe Geo-Institute. However, these foreign programs are difficult to interpret when applied to a country such as Korea where there are many mountainous terrains, and have problems due to high price.

SLOPILE for windows developed by ENG Engineering & Engineering Co., Ltd. and Civil Engineering Consulting Co., Ltd. were developed to release these problems using anchors, piles and geotextiles. These programs are highly reliable slope stability analysis programs that fit well with the domestic environment.

However, in general, a disaster such as a slope collapse is a disaster caused by a long period of continuous stimulation, and a large number of detection sensors must be installed to improve reliability for preventing a disaster. Therefore, there is a disadvantage in that excessive communication costs are required because the data input from the plurality of sensing sensors are transmitted to the server and monitored in real time.

The present invention was devised to solve such a problem, and an object thereof is to provide a slope disaster monitoring system that analyzes measurement data in a field and transmits detection data in real time when an abnormal situation occurs.

Furthermore, the present invention provides a slope disaster monitoring system that compresses and transmits sensing data every predetermined period when an abnormal situation does not occur.

In addition, the present invention provides a slope disaster monitoring system that provides detection information of a corresponding area to a large number of users regardless of time and place through the web.

Disaster detection device of the slope disaster monitoring system according to an aspect of the present invention for achieving the above object receives the sensing data input from a plurality of detection sensors installed in the disaster occurrence area. The received sensing data determines whether the abnormal data, that is, the data with a high probability of occurrence of abnormal symptoms, by the automatic computation neural network. If it is not the abnormal data, the automatic computational neural network compresses the data and stores it in memory. The sense data compressed and stored in the memory is transmitted to the monitoring server at predetermined intervals set by the administrator through the wireless modem.

The monitoring server receives compressed data from a plurality of disaster detection devices and restores the compressed data to a disaster information database. The data stored in the disaster information database is used for analysis, monitoring, and web services of the measurement area.

Meanwhile, when the sensing data input from the plurality of sensing sensors is data of an abnormal pattern, the disaster sensing apparatus transmits the sensing data to the monitoring server without compressing the sensing data in real time regardless of a predetermined cycle. The monitoring server receives the uncompressed sensed data, generates an alert sound, for example, a beat sound or voice guidance, to the administrator, and allows the administrator to analyze the received sensed data.

According to an additional aspect of the present invention, the monitoring server of the slope disaster monitoring system according to the present invention may provide information of a corresponding measurement area in which a disaster detection device is installed to a plurality of users in real time through a network such as the Internet. The monitoring server according to the present invention further includes a web service providing unit capable of supporting information of a corresponding measurement area in real time as a user terminal connected to the Internet or a network and connected to the server. Therefore, the user can receive information of the corresponding measurement area in real time from the monitoring server regardless of time and place using his terminal, that is, a computer connected to the Internet or a network.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily understand and reproduce the present invention.

1 is a general schematic diagram of a slope disaster monitoring system according to an embodiment of the present invention. As shown, the slope disaster monitoring system according to the present invention comprises a plurality of disaster detection apparatus 100, a plurality of user terminal 300 connected to the monitoring server 200 and the network such as the Internet.

The disaster detection apparatus 100 receives detection data input from the detection sensor 110 installed in a region where a disaster is likely to occur and determines whether the detection data is an abnormal pattern. Compress and save The disaster detection apparatus 100 transmits the compressed data stored at every predetermined cycle input by the administrator to the monitoring server 200 through the wireless communication modem. In addition, in the case of the detection data of the abnormal pattern, that is, the detection data having a high probability of occurrence of abnormal symptoms, the detection data is transmitted to the monitoring server 200 in real time regardless of the transmission period set by the administrator without compressing the corresponding data.

The monitoring server 200 receives and restores the compressed sensing data from the disaster sensing apparatus 100, and notifies the administrator that an abnormal symptom has occurred when receiving the abnormal pattern of sensing data, that is, the uncompressed sensing data. In addition, the monitoring server 200 provides detection information obtained by analyzing detection data transmitted from the disaster detection apparatus 100 according to a request of a plurality of computers connected to the Internet or a network.

2 is a schematic block diagram of a sensing device according to an embodiment of the present invention. As shown, the disaster detection apparatus 100 of the slope disaster monitoring system according to the present invention includes a plurality of detection sensors 110, a memory 130, a switching unit 140, a data transmitter 150, The abnormal signal detection unit 160 is included.

A plurality of detection sensors 110 are installed in a region where a high possibility of a disaster occurs, and detects each unique sensing value and outputs it to the disaster detection apparatus 100. According to an embodiment of the present invention, the detection sensor 110 connected to the disaster detection apparatus 100 includes three modules, a rainfall sensor for measuring rainfall, a TW (Temperature Wire) sensor, and a VW (Vibrating Wire) sensor. The module is connected to the disaster detection device 100 by RS-485. The sensing data sensed by the sensing sensor 110 is converted into digital by an analog-to-digital converter and transmitted to the automatic computation neural network 120 of the disaster detection apparatus 100 connected in an RS-485 manner. Up to 64 sensors may be connected to one disaster detection apparatus 100.

The memory 130 stores sensing data output from the sensing sensor 110. The switching unit 140 receives one of the compressed data read out from the memory 130 and the sensing data directly input from the sensing sensor 110 according to the control signal output from the abnormal signal detecting unit 160 and is selectively selected. Will print The data transmitter 150 is, for example, a communication modem such as a wireless communication modem, and transmits sensing data selectively output by the switching unit 140 to the monitoring server 200.

The abnormal signal detection unit 160 reads and outputs the sensing data stored in the memory 130 at predetermined intervals when the sensing data input from the sensing sensor 110 is less than or equal to a prescribed value, for example, when there is no slight change in the slope. Outputs a control signal of the switching unit 140 to. In addition, when the detection data is more than a prescribed value, for example, when the detection data that is rapidly changed from the detection sensor 110 installed on the slope is detected and received by the disaster detection apparatus 100 in real time monitoring the corresponding detection data ( The control signal of the switching unit 140 is output to be transmitted to the 200.

According to a further aspect of the present invention, the disaster detection apparatus according to the embodiment of the present invention further includes an automatic operation neural network 120. The automatic computation neural network 120 compresses the sensing data input from the sensing sensor, learns the input sensing data, and outputs the received sensing data. The phase signal detector 160 detects an abnormality of the corresponding data by using the residual of the sensed data input from the sensor 110 and the learned sensed data output from the automatic computation neural network, and accordingly, the switching unit 140. Outputs a control signal. The abnormal signal detection unit 160 estimates the detection data of the normal pattern when inputting the detection data such as the change of the detection data occurring before the collapse of the slope, for example, a pattern different from the data used in the learning. The threshold check of the residuals obtained through 'enables the detection timing of abnormal signs.

When the threshold value of the residual is less than the prescribed value based on the detection time of the abnormal symptom, the data stored by the automatic operation neural network 120 is stored in the memory 130 by the switching unit 140, and the monitoring server (for example) The control signal of the switching unit 140 is output to be transmitted to the 200. If the threshold value is equal to or greater than a predetermined value, that is, when an abnormality occurs, the control signal of the switching unit 140 is transmitted to the monitoring server 200 in real time without compressing the sensing data input from the sensing sensor instead of the compressed sensing data. Output

3 is a schematic block diagram of a layer structure of an automatic computation neural network of a disaster detection apparatus according to an embodiment of the present invention. As shown, the automatic computational neural network 120 of the disaster detection apparatus 100 according to the present invention includes an input layer 121, a mapping layer 122, and a bottleneck layer. A layer 123, a demapping layer 124, and an output layer 125.

The input layer 121 receives M-dimensional data X ₁ ... X _M as an input. The input sensing data are sensor values measured from one disaster detection apparatus 100 and they are related data. The mapping layer 122 is composed of L-dimensional neurons which perform a function of reducing M-dimensional sense data input from the input layer 121 to fewer F-dimensional feature data than its input / output characteristics. It is represented by Equation 1.

In Equation 1, a vector having L-dimensional neuron activation functions (G ₁ , G ₂ , ..., G _L ) of the mapping layer 122 as elements, X denotes a measurement variable of 1xM dimension, and F is Refers to abbreviated data of 1xF dimension. This characteristic enables compression from the M-dimensional sense data to the F-dimensional data, and the compressed data can be obtained through the output of the Bartlek layer 123.

The demapping layer 124 is composed of L-neurons that perform the function of restoring the reduced 1xF-dimensional data to the original dimension (1xM), which is represented by Equation 2 below.

In Equation 2, a vector having L-dimensional neuronal activity functions (H ₁ , H ₂ , ..., H _L ) of the demapping layer 124 as elements and Y is the same as that used as an input of an autocorrelation neural network. 1xM dimensional output data. The neuronal activity functions G and H of Equations 1 and 2 include coupling strengths to be adjusted during learning, and they are trained to minimize the residuals according to Equation 3 by a general teacher learning method.

4 is a schematic block diagram of a monitoring server according to an embodiment of the present invention. As shown, the monitoring server 200 of the slope disaster monitoring system according to the present invention includes a data receiver 210, a decoder 230, a disaster information database 250, an image output unit 260, The voice output unit 270 and the disaster signal detection unit 220 is included.

The data receiver 210 is, for example, a communication modem such as a wireless communication modem, and receives the sensed data transmitted from the disaster detection apparatus 100 and outputs it to the disaster signal detection unit 220.

The disaster signal detector 220 receives sensing data input from the data receiver 210, determines whether the sensed data is compressed data, and selectively outputs the sensed data. If the received sensed data is compressed data, it is output to the decoder 230. In addition, when the sensed data of the uncompressed abnormal pattern is input, the uncompressed sensed data is output to the image output unit 260 and displayed to the administrator, and an abnormal situation occurs to the administrator by outputting a driving signal to the audio output unit 270. Notify you that

The disaster information database 250 stores the sensed data decrypted by the decryption unit 230. The image output unit 260 outputs the decoded sensed data or the uncompressed sensed data, that is, the sensed data of the abnormal pattern, output from the decoder 230 to the manager. The voice output unit 270 notifies the system administrator in the form of a beat sound or voice announcement when a disaster occurs.

5 is a schematic block diagram of a monitoring server according to another embodiment of the present invention. As shown, according to a characteristic aspect of the present invention, the decryption unit 230 of the monitoring server 200 of the slope disaster monitoring system according to the present invention may be implemented as a non-linear autonomic neural network 240. The automatic computation neural network 240 decodes the compressed sensed data output by the disaster signal detection unit 220 and stores the decoded sensed data in the disaster information database 250. Detailed descriptions of the non-linear autonomic neural network 240 will be described with reference to FIG. 6. 6 is a schematic block diagram of an automatic computational neural network of a nonlinear technique in accordance with an embodiment of the present invention. The automatic computational neural network 240 according to an embodiment of the present invention has a 1: 1 mapping function, but when used separately, efficient compression and restoration of multivariable data having a nonlinear relationship is possible. As shown in FIG. 6, the neural network having the Bartlek layer 241 as the output terminal is shortened and output the M-dimensional data to the F-dimensional data at the first half, and the Neural network having the Bartlek layer 241 as the input terminal is F in the second half. It performs the function of restoring the -dimensional data to the M-dimensional data.

Accordingly, the monitoring server 200 of the slope disaster monitoring system according to the present invention has an automatic computing neural network having a nonlinear relationship including only the demapping layer 242 and the output layer 243 as the input of the Bartle layer 241. By including the 240, effective data recovery may be performed on the compressed sensed data transmitted from the disaster detection apparatus 100.

According to a further aspect of the present invention, the monitoring server 200 of the slope disaster monitoring system according to the present invention includes a web service provider 280. The web service provider 280 transmits the corresponding sensed data, that is, the data analyzed by the monitoring server 200, to an ActiveX control driven in the plurality of user terminals 300. The web service provider 280 is connected to the user terminal 300 by a TCP / IP interface. Data transmission methods using the Internet include TCP (Transmission Control Protocol) and UDP (User Datagram Protocol). UDP is advantageous for real time transmission in data transmission but does not guarantee data integrity. When real time transmission such as internet broadcasting or internet phone is important, UDP method is used, but TCP is used when connection setting, flow control, and error control are required. In the present invention, the user terminal 300 connected to the monitoring server 200 is used. TCP was used for data transfer between

When the user connects to the domain of the IP specified in the web service provider 280 of the monitoring server 200 through a web browser using his computer, the web service provider 280 is an application object ( Application Object) and a Session Object are created, and information provided to the user terminal 300 is provided, that is, HTML, and then waits for connection of another user terminal 300. Through such an operation, the web service provider 280 has a multi-user environment in which one monitoring server 200 is connected to multiple user terminals 300 in a multiple.

The web service provider 280 includes a CAB file for using ActiveX in addition to the HTML file, and the ActiveX operates on the web browser by the "CLASSID" value. If the corresponding "CLASSID" does not exist in the user terminal 300 connected to the web service provider 280 of the monitoring server 200, the CAB file is automatically downloaded and setup.

As described above, the slope disaster monitoring system according to the present invention determines whether there is an abnormality of sensing data inputted by a plurality of disaster detection devices, and transmits the corresponding data to the monitoring server in real time when data of an abnormal pattern is input, thereby generating a disaster. Early detection has a quick suppression effect.

In addition, when the input sensing data is not the data of the abnormal pattern, the data is compressed and stored, and the compressed data stored in each predetermined cycle is read and transmitted, thereby reducing the amount of data transmitted and efficient data transmission. .

In addition, the monitoring server can provide the sensing data transmitted from the disaster detection device to a plurality of user terminals through the web, and the administrator can monitor the measurement area regardless of the place using a computer connected to the Internet or a network. Has an effect.

The present invention has been described above with reference to preferred embodiments, but is not limited thereto, and is interpreted by the appended claims, which are intended to cover many modifications that will be apparent to those skilled in the art without departing from the scope of the present invention. Should be done.

Claims (7)

delete In the disaster detection device for receiving the disaster detection data input from a plurality of detection sensors and transmitting to the monitoring server, the device: A memory for storing sensing data output from the sensing sensor; An input layer for receiving multidimensionally related sensing data from the sensing sensor, a mapping layer for compressing the multidimensional sensing data output from the input layer into feature data having a smaller dimension, and the feature data compressed by the mapping layer. Auto-neural network including a Bartlek layer for storing the data in the memory, a demapping layer for reconstructing the feature data compressed by the mapping layer, and an output layer for outputting the sensed data reconstructed by the demapping layer. ; A switching unit for selectively outputting one of compressed data read out of the memory and uncompressed data input from the sensing sensor every predetermined period; A data transmitter for wirelessly transmitting data output by the switching unit to the monitoring server; The control signal of the switching unit is read out to output the compressed data stored in the memory every predetermined period when the threshold value is equal to or less than a prescribed value through the residual threshold check of the sensing data input from the sensing sensor and the data output from the output layer. Output, An abnormal signal detection unit outputting a control signal of the switching unit to transmit to the monitoring server in real time without compressing the corresponding sensing data when the threshold value is equal to or greater than a prescribed value; Disaster detection device comprising a. delete A monitoring server for receiving data from a disaster detection device that receives sensing data from a plurality of sensing sensors and selectively compresses and transmits the sensing data, the monitoring server comprising: A data receiving unit which receives sensing data transmitted from the disaster detection device; Outputs a bartle layer for receiving compressed data output from the disaster detection device, a demapping layer for reconstructing the compressed data output from the bartle layer into data upon input, and sensing data reconstructed by the demapping layer A decoder including an automatic computation neural network having a nonlinear relationship including an output layer; An image output unit which displays the sensed data output from the data receiver or the sensed data decoded by the decoder; A voice output unit which notifies the system administrator of the occurrence of a disaster through a beat tone or a voice announcement when a disaster occurrence signal is received; A disaster information database for storing the sensed data decoded by the decoder or the uncompressed sensed data output from the data receiver; Determining whether the sensed data output from the data receiver is compressed data and outputting compressed data to the decoder in the case of compressed data; A disaster signal detector for outputting a disaster occurrence signal to the voice output unit in case of uncompressed data; Disaster detection monitoring server comprising a. The method of claim 4 wherein the server is: A web service providing unit configured to read and provide detection data of a corresponding disaster area from the disaster information database according to a request of a plurality of user terminals connected through the Internet or a network; Disaster detection monitoring server characterized in that it further comprises. delete A disaster monitoring system comprising a disaster detection device receiving a disaster detection signal input from a plurality of detection sensors and transmitting a disaster occurrence signal, and a monitoring server receiving a disaster detection signal from the disaster detection device to notify an administrator of the disaster detection device, wherein Disaster Detection Devices: A memory for storing sensing data output from the sensing sensor; An input layer for receiving multidimensionally related sensing data from the sensing sensor, a mapping layer for compressing the multidimensional sensing data output from the input layer into feature data having a smaller dimension, and the feature data compressed by the mapping layer. Auto-neural network including a Bartlek layer for storing the data in the memory, a demapping layer for reconstructing the feature data compressed by the mapping layer, and an output layer for outputting the sensed data reconstructed by the demapping layer. ; A switching unit for selectively outputting one of compressed data read out of the memory and sensing data input in real time from the sensing sensor every predetermined period; A data transmitter for wirelessly transmitting data output by the switching unit to the monitoring server; The control signal of the switching unit is read out to output the compressed data stored in the memory every predetermined period when the threshold value is equal to or less than a prescribed value through the residual threshold check of the sensing data input from the sensing sensor and the data output from the output layer. Output, An abnormal signal detection unit for outputting a control signal of the switching unit to transmit to the monitoring server in real time without compressing the corresponding sensing data when the threshold value is equal to or greater than a prescribed value; The disaster monitoring server: A data receiver configured to receive compressed data or sensing data transmitted from the disaster detection apparatus; A bartlek layer for receiving the compressed data output from the data receiver, a demapping layer for reconstructing the compressed data output from the bartlek layer to data upon input, and the sensed data reconstructed by the demapping layer; A decoder including an automatic computation neural network having a nonlinear relationship including an output layer; An image output unit which displays the sensed data output from the data receiver or the sensed data decoded by the decoder; A voice output unit which notifies the system administrator of the occurrence of a disaster through a beat tone or a voice announcement when a disaster occurrence signal is received; A disaster information database for storing the sensed data decoded by the decoder or the uncompressed sensed data output from the data receiver; A web service providing unit configured to read and provide detection data of a corresponding disaster area from the disaster information database according to a request of a plurality of user terminals connected through the Internet or a network; Determining whether the sensed data output from the data receiver is compressed data and outputting compressed data to the decoder in the case of compressed data; A disaster signal detector for outputting a disaster occurrence signal to the voice output unit in case of uncompressed data; Slope disaster monitoring system characterized in that.

Images