JPH06282780A - Wireless security device - Google Patents

Abstract

PURPOSE:To provide a wireless security device which improves the noise resistance. CONSTITUTION:The wireless security device includes a frequency analyzing means which divies the signals equivalent to one cycle of the signal that can be extracted from a section where an answer signal should be included, a neural network 9 which is connected to the frequency analyzing means through the input layer of the network and previously learnt so as to output through an output layer the affirmative value when the signals including an answer signal are supplied through the input layer and to output the negative value when only the noises are supplied through the input layer respectively, and a deciding means 10 which decides the presence of absence of an answer signal by the output value of the network 9. In such a constitution, a state monitoring part 1 decides the presence or absence of the answer signal received from a sensor part 2.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a wireless security device for managing the safety of a building such as the locked state of doors and windows, the open / closed state of gas taps, the connected state of electric appliances and the like.

[0002]

2. Description of the Related Art Conventionally, a wireless security device of this type is known from Japanese Patent Application Laid-Open No. 3-118691. This structure is shown in FIG. This wireless security device is composed of a state monitoring unit 1 for detecting the state of a monitored object and a sensor unit 2 attached to the monitored object. The state monitoring unit 1 generates a monitoring signal, a signal generating unit 15, a transmitting / receiving antenna 6 that emits the signal to the sensor unit 2 and receives a signal including a response signal from the sensor unit 2, and a receiving unit. Signal monitoring means 16 for monitoring the section in which the response signal in the generated signal should exist
And a control means 18 for controlling the operation of the transmission / reception changeover switch 17 and the entire apparatus. Further, the sensor unit 2 includes a transmission / reception antenna 12 for exchanging signals with the state monitoring unit 1, and a switch 13 that is turned on / off depending on the state of the monitored object, and resonates with the frequency of the monitoring signal. Forming a circuit.

In such a configuration, the state monitoring unit 1
However, when a signal for monitoring is transmitted from the transmitting / receiving antenna 6, the sensor unit 2 is for monitoring when the switch 13 of the sensor unit 2 is turned on and the circuit is closed due to the mutual induction action and resonance action of the electromagnetic coupling. A response signal will be issued immediately after the reception of the signal is completed. Therefore, the state monitoring unit 1 extracts the section in which the response signal of the sensor unit 2 should exist from the received signal by the signal monitoring unit 16 and then identifies the presence or absence of the signal level or the frequency in the section. The presence or absence of a response signal generated by the sensor unit is determined, and the state of the monitored object is detected based on the result.

However, in such a conventional wireless security device, the presence or absence of a response signal generated by the sensor unit is determined by identifying the presence or absence of the signal level or the frequency in a predetermined section of the signal received by the state monitoring unit. Therefore, when noise is received or noise is superimposed on the response signal, there is a problem that an erroneous determination is likely to be made.

[0005]

SUMMARY OF THE INVENTION The present invention has been made in view of the above problems, and an object of the present invention is to provide a wireless security device having improved noise resistance performance.

[0006]

Therefore, in the present invention, in particular, in the state monitoring section, the section for judging the presence or absence of the response signal generated by the sensor section is configured as follows. That is, a frequency analysis means for dividing a signal for one period of a signal formed by extracting a section in which a response signal should exist from a received signal into a plurality of frequency bands, and an input layer connected to this frequency analysis means, and a response signal A neural network that is pre-learned so that a positive value is output when a signal including is input from the input layer, and a negative value is output from the output layer when only noise is input from the input layer, and the output value of this neural network. A determination means for determining the presence or absence of the response signal is provided.

[0007]

With the above configuration, a signal for one period of the signal in the section where the response signal should exist is taken out, and the output value of each band obtained by dividing the signal into a plurality of frequency bands is regarded as one pattern. However, by using the neural network learned in advance, it becomes possible to determine whether or not the signal received by the state monitoring unit includes the response signal.

[0008]

EXAMPLES The present invention will be described in detail below based on examples. FIG. 1 is a block diagram showing an embodiment of a wireless security device according to the present invention. In the figure, the wireless security device comprises a state monitoring unit 1 for detecting the state of a monitored object and a sensor unit 2 attached to the monitored object. The state monitoring unit 1 includes an original oscillator 3, a frequency divider 4 that reduces the frequency to 1 / n, and a burst wave generator 5 that generates a burst wave from the outputs of the original oscillator 3 and the frequency divider 4. The monitoring signal is generated by these. The state monitoring unit 1 also has a transmission / reception antenna 6, and the output of the burst wave generator 5 is connected to the transmission side. Further, a detector 7 connected to the receiving side of the transmitting / receiving antenna 6, a signal processor 8 for extracting a predetermined output from the detector 7, a neural network 9 having an input layer connected to the signal processor 8, and an output layer of the neural network 9. It has a connected determiner 10 and a display 11 for displaying the state of the monitored object based on the determination result. On the other hand, the sensor unit 2 has a transmission / reception antenna 12, a switch 13 that is turned on / off depending on the state of the monitored object, and a solid-state resonator 14, so that the frequency of the monitoring signal emitted by the state monitoring unit 1 can be adjusted. It forms a resonance circuit that resonates with the resonance circuit. A capacitor may be included in this resonance circuit. Further, a crystal resonator, a ceramic resonator, or the like may be used as the solid resonator.

A usage mode of the above configuration will be described. The state monitoring unit 1 is installed at a specific place in the building, and the sensor unit 2 is attached to a door, a window, a gas tap, an electric appliance, or the like. When the sensor unit 2 is attached to a door or a window, if the switch 13 and the lock are linked, the state monitoring unit 1 can detect the locked state depending on the presence or absence of a response signal. The open / closed state of the gas tap can also be detected in the same manner if the switch 13 and the tap are linked. The same applies to the connection state of electric appliances.

FIG. 2 is a diagram showing the output signal of each means in FIG. The operation of the embodiment will be described with reference to FIG. 1 while referring to the configuration of FIG. First, the transmission operation from the state monitoring unit 1 to the sensor unit 2 will be described. The original oscillator 3 is a circuit that generates a signal having a specific frequency f.
A pulse wave is generated like the signal (A). The frequency divider 4 converts the frequency f of the output signal of the original oscillator 3 into 1 / n, and the output is as shown in signal (B). The burst wave generator 5 generates a signal (C) from the signal (A) output from the original oscillator 3 and the signal (B) output from the frequency divider 4. Then, this signal (C) is transmitted from the transmitting / receiving antenna 6 to the sensor unit 2 as a monitoring signal.

Next, the receiving operation and the transmitting operation of the sensor section 2 will be described. The monitoring signal and the signal (C) transmitted from the state monitoring unit 1 are received by the transmission / reception antenna 12. Specifically, a current of frequency f is intermittently applied to the transmission / reception antenna 6 of the state monitoring unit 1, and an electromotive force is generated in the transmission / reception antenna 12 of the sensor unit 2 by the mutual induction action of electromagnetic coupling. When the switch 13 of the sensor unit 2 is turned on, the transmitting / receiving antenna 12,
Since the resonance circuit is formed so as to resonate at the frequency f with the solid resonator 14, the maximum current is generated. Energy is temporarily stored in the solid state resonator 14 by this current, and is released when the signal from the state monitoring unit 1 is stopped to generate a response signal. That is, a current flows in the transmission / reception antenna 12, an electromotive force is generated in the transmission / reception antenna 6 of the state monitoring unit 1 by the mutual induction action of the same electromagnetic coupling as described above, and a signal is transmitted. If the sensor unit 2
When the switch 13 of is off, it does not resonate,
No response signal is generated.

Next, the operation from the receiving operation of the status monitor 1 to the status determination will be described. The signal received from the transmission / reception antenna 6 is superposed on the monitoring signal transmitted by itself when the response signal is present, so that the signal (D)
become that way. Further, when there is no response signal and only noise is received, the signal (D ′) is obtained. The detector 7 extracts the section in which the response signal should exist from the signal (D) or the signal (D ′) according to the transmission time, excluding the portion of the monitoring signal. The output obtained by squaring the signal existing in the extracted section is the signal (E) or the signal (E ′). The band-pass filter 8 divides the signal (E) or the signal (E ′) for one cycle into a plurality of frequency bands. Then, the output value of each band of the bandpass filter 8 is regarded as one pattern, and the neural network 9 discriminates whether it is a response signal or noise. In particular, this process will be described with reference to FIG.

FIG. 3 is a diagram for explaining the determination process using a neural network. Neural network 9 input layer a plurality of units, the output layer is composed of one unit, e _1, e ₂ is the output value of each band of the band-pass filter, ..., the e _n, respectively neural network Input is made in correspondence with each unit of the 9 input layers. The neural network 9 uses e ₁ ,
When e ₂ , ..., E _n are signals including a response signal,
Positive value, for example, "1", e _1, e _2, ..., are trained to when e _n is a signal in the noise only output from the unit of the output layer negative value, for example, a "0" . The determiner 10 has a preset threshold value, and compares this value with the output value of the neural network 9 to determine whether it is a signal including a response signal or a signal containing only noise.
Here, the output layer of the neural network is set to 1
Although an example is shown that it is composed of two units, it is composed of two units,
You may make it allocate the determination output of the signal containing only a noise and the signal containing a response signal to each unit. Also,
The threshold value set in the judging device may be set in two stages, and a rejection (decision impossible) result output may be provided.

Next, the learning method will be described with reference to FIG. FIG. 4 (X) is a set of output values of the band pass filter for one cycle of the signal including the response signal. This is created by collecting a plurality of patterns which are deformed by superimposing various noises on the response signal. The set (Y) is a set of output values of the bandpass filter created only from noise. Using these data, an affirmative value “1” is output when the data of the set (X) is input to the neural network, and a negative value “0” is output when the data of the set (Y) is input. Is trained by the error back propagation method.

Another embodiment relating to the determination processing part is shown in FIG.
It was shown to. In the above-described embodiment, the response value or the noise is discriminated using the output value of the bandpass filter for one cycle, but in the present embodiment, the bandpass filter for several cycles is used as shown in FIG. Judgment is performed using the output value of the filter. The output value of the bandpass filter for several cycles can be created by extracting from the signal (E) or the signal (E ′). Also,
By transmitting the monitoring signal a plurality of times, it is also possible to use the data of the reception signal obtained respectively. That is, each data (1), (2), ..., (N) is input to the neural network 9 and individually determined,
A final decision is made based on the number of results. For example, when the result that the signal includes the response signal is output K times or more, it is determined that the signal is the response signal. However, the determination method may be a majority vote, and if the result that the noise is generated is output even once from the viewpoint of safety, it may be determined that the noise. As described above, when the determination is performed based on the output value of the band pass filter for several cycles, it is possible to deal with sudden noise and the determination accuracy is further stabilized.

[0016]

As described above in detail, according to the present invention,
Even in an environment where noise is likely to be superimposed on the signal, it is possible to provide a wireless security device that can maintain a long communication distance between the state monitoring unit and the sensor unit without lowering the accuracy of determining the presence or absence of the response signal emitted by the sensor unit.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of a wireless security device according to the present invention.

FIG. 2 is an output waveform diagram of each unit in FIG.

FIG. 3 is a diagram illustrating a determination process using a neural network.

FIG. 4 is a diagram for explaining a learning method of a neural network.

FIG. 5 is a diagram illustrating another embodiment of the determination process using a neural network.

FIG. 6 is a block diagram showing a configuration of a conventional wireless security device.

[Explanation of symbols]

 1 state monitoring unit 2 sensor unit 6, 12 transmission / reception antenna 8 band pass filter 9 neural network 10 decision device 13 switch

Claims (2)

[Claims] 1. Signal generation means for generating a monitoring signal,
A transmitting / receiving antenna that emits the monitoring signal toward a sensor unit attached to a monitored object and receives a signal including a response signal from the sensor unit, and extracts a section in which the response signal should exist from the received signal. It includes a state monitoring unit composed of signal extracting means, a transmission / reception antenna for exchanging signals with the state monitoring unit, and a switch that is turned on / off depending on the state of the monitored object, and resonates with respect to the frequency of the monitoring signal. In a wireless security device having a sensor unit forming a resonance circuit, a frequency analysis unit that divides a signal of one cycle of an output signal of the signal extraction unit into a plurality of frequency bands, and an input layer is connected to the frequency analysis unit. , When the signal including the response signal is input from the input layer, the positive value, when only noise is input from the input layer, the negative value is output layer Wireless security device comprising a neural network pre-trained to al outputs, in that a determination means for determining whether said response signal by the output value of the neural network. 2. The determining means makes a plurality of determinations using the output of the frequency analyzing means for several cycles, and finally determines the presence or absence of the response signal based on the result. The wireless security device according to claim 1.