JPS6211998A - Voice alarm - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an alarm device for disaster prevention equipment, etc., and more particularly to an alarm device that synthesizes a warning message and issues the warning message in the form of voice.

[Prior Art and Problems to be Solved by the Invention] Various facilities in buildings such as factories and buildings include, for example, elevator facilities, electric power facilities, gas facilities, and air conditioning facilities. When a failure or accident occurs in these facilities, it is necessary to constantly monitor the facilities and issue an alarm as necessary in order to minimize damage. by the way,
Conventionally, such warnings have been typically issued using sounds such as buzzers, bells, sirens, etc., and/or lamp displays. However, among the above-mentioned buildings, there are many types of alarms and facilities where alarm devices are installed in many places, and in some cases, IT may not be possible to detect an abnormal situation with just such sounds and lamp displays. Even if the location of the occurrence is known, the content and extent of the abnormal situation cannot be quickly ascertained, and therefore, it is expected that it will take time to take appropriate measures, resulting in unexpected damage. In view of these circumstances, instead of notifying the police informant with a simple alarm sound,
As in Japanese Patent Application No. 152096, it has been proposed that a warning message is synthesized into a voice and notified using a concrete voice. In this type of conventional example, a number of predetermined warning messages are stored in a ROM (read-only memory) and read out from the ROM as needed.

In such a ROM, once a warning message is stored, it is not easy to change the stored contents of the warning message. On the other hand, a predicted warning message alone may not be enough to deal with the situation more accurately. In such cases, the conventional ROM storage method, which does not allow the alarm message to be replaced (changed or added), is extremely inconvenient, and replacing the alarm message requires considerable cost and time. It had become. In addition, in conventional speech synthesis, for example, Pacor (PA
This method uses an analysis and synthesis method such as the RCOR method, which requires a lot of money and time for voice analysis, and has the disadvantage that users cannot easily register new warning messages as necessary.

The present invention has been made in view of the above circumstances, and includes:
To provide a voice warning device capable of quickly and easily analyzing voice and easily rewriting a warning message. In addition, in the present invention, the alarm message is
This includes, but is not limited to, words and phrases that inform the type, location, and severity of an abnormal situation, as well as words and phrases that describe emergency exit guidance, guidance on how to operate machinery in the event of an accident, etc.

[Means for solving problems]

Hereinafter, the present invention will be explained in detail based on embodiments shown in the drawings. FIG. 1 is a block circuit diagram of a main part of an alarm device in disaster prevention equipment, etc. according to an embodiment of the present invention. Code 1 is
This is a sensor unit that is placed at a predetermined location and outputs a data signal corresponding to an abnormal situation at the location.This sensor unit includes, for example, analog sensors 1a to 1n that measure temperature, humidity, air flow rate, etc. detectors, gas detectors, people,
It includes digital sensors 1a to 1n such as object passage detectors.

In the case of an analog sensor, the sensor output is subjected to A/D conversion processing. Reference numeral 2 denotes an abnormal situation determination unit that determines whether or not the surrounding situation is abnormal in response to sensor outputs from each sensor in the sensor unit 1, and 3 an alarm message memory that stores an alarm message corresponding to the abnormal situation. The unit 4 stores an alarm message storage unit 3 based on the judgment output of the abnormal situation judgment unit 2.
5 is a speech synthesis section that performs speech synthesis based on the alarm message output of the alarm message selection section 4; 6;
Reference numeral 7 denotes a noise removal unit that removes noise such as granular noise contained in the output of the speech synthesis unit 5; 7 is an audio output unit that outputs audio according to the output of the noise removal unit 6; is composed of, for example, a plurality of speakers 7a to 7n individually corresponding to each sensor of the sensor section 1. Reference numeral 7 denotes an input section such as a microphone for setting an alarm message.

The alarm message storage section 3 is composed of a RAM that can be written into and read out at any time, and the speech synthesis section 5 includes an LSI for speech synthesis using a waveform encoding method.

[Effect]

Since the alarm message storage section 3 is composed of a RAM, it can be stored and read out at any time, and the audio input from the alarm message setting input section 8 is used for speech synthesis using the waveform encoding method included in the speech synthesis section 5. It is easily digitized by an LSI and stored in the RAM, and necessary alarm messages are synthesized according to the output signal of the alarm message selection section 4.

Further, noise such as granular noise, which is a problem in training synthesis using the waveform encoding method, is removed by the noise removal unit 6.

〔Example〕

Figure $2 shows details of the alarm message storage section 3, speech synthesis section 5, noise removal section 6, and alarm message input section 8 in one embodiment of the present invention.

In FIG. 2, 51 is ADM (Adaptive
LSI for speech synthesis using the eDelta Modulation method (Toshiba T 6831)
L s ■) Te. Here, the ADM method is one of the waveform encoding methods. 52 is a controller that controls the generation timing of an alarm message, for example;
This controller 52 provides the generation timing signal to the speech synthesis r-si 51. T-5I for speech synthesis
51 controls the generation timing of the alarm message in response to this generation timing signal.

As already mentioned, the alarm message storage section 3 includes a RAM for rewriting the alarm message as shown in FIG.
(write/read memory at any time), and in such a configuration, the voice synthesis LSI 51
By storing the alarm message set by the alarm message setting input section 8 in the alarm message rewriting RAM 3, it is possible to add or change the alarm message.

Although the power supply is not shown in the above embodiment, it is preferable that the power supply includes, for example, a backup power supply for the RAM 3 and a power failure detection circuit. That is, when the main power source is cut off or there is a power outage, the contents written in the RAM 3 are prevented from being destroyed by the backup power source.

The noise removal section 6 includes, for example, a low-pass filter 61, a lower limiter circuit 62, a bandpass filter 63, and an output amplifier 64. The low-pass filter 61 is provided for impedance conversion. At the lower end, the vine circuit 62 is configured as shown in FIG. 3, for example. 621A, 622A, 623
A is operational amplifier, 621D, 621D2, 622D,
, 622D2 is a diode, 621R1~621k
a, 622”1 to 622Rn.

623kl~623Ra, 624R1, 624R
2 is a resistor, and 625 is a capacitor.

By applying a positive voltage E to the terminals 626 and 627 and adjusting the voltage adjustment variable resistors 621R, 622R□,
Appropriate bias voltages E' and E are obtained from contacts 628 and 629, respectively. Only when the signal voltage e input from the input terminal 630 is e)E' or e<E'', the diodes 621D□ and 622D1 conduct, respectively, and the capacitor 625, resistor 621R, 622R4 operational amplifier 621A, 622A, etc. through the resistor 62
The signals are added by the operational amplifier 623A via 1R6 and 622R6, and output to the terminal 631. 621D2゜622
D2 is a diode for preventing the operational amplifiers 621A and 622A from being saturated when 621D1 and 6221)l are OFF. Since the lower limiter circuit 62 is configured as described above, it is possible to prevent noise such as granular noise in the audio output from the audio synthesizer 5 from passing through.

The bandpass filter 63 passes a frequency range of 500 to 4,000 H2, which is a frequency range that humans can easily perceive as sound, and removes waveform distortion caused by DA conversion. Needless to say, the output amplifier 64 is for adjusting the volume. The audio input section 8 includes a microphone 81, an input volume adjustment amplifier 82, and a low-pass filter 83. The low-pass filter 83 is, for example, 4,00
It is provided to cut off frequency components on OH2Ja and compress the amount of information. Sensor part 1. Abnormal situation determination section 2° Alarm message selection section 4. The audio output section 7 is as described based on FIG.

〔Effect of the invention〕

As described above, the audio alarm device of the present invention has an input section for setting an alarm message and a noise removal section, the speech synthesis section includes an LSI for speech synthesis using a waveform encoding method, and the alarm message storage section has an input section for setting an alarm message and a noise removal section. Since it is composed of a memory RAM, the voice synthesis LS is inputted from the alarm message setting input section to the alarm message memory RAM.
Any alarm message can be read out at any time via I, and the cost and time required for replacing the message can be reduced. Furthermore, since the noise removing section is provided, high quality audio output can be obtained in which noise such as granular noise, which is a problem in audio synthesis using the waveform encoding method, is removed.

[Brief explanation of drawings]

1 is a sensor section, 2 is an abnormal situation judgment section, 3 is an alarm message storage section (RAM), 4 is an alarm message selection section, 5 is a speech synthesis section, 6 is a noise removal section, 7 is an audio output section, and 8 is an alarm Message setting input section, 51 is a voice synthesis LSI,
52 is a lower limiter circuit. Patent applicant Tak Electric Cable Co., Ltd. Agent Kama 1) Bun 2

Claims (2)

[Claims] (1) A sensor unit that is placed at a predetermined location and outputs a data signal corresponding to the abnormal situation at that location, and based on various data signals from this sensor unit, the situation around the sensor unit is abnormal. an abnormal situation determining unit that determines whether or not the abnormal situation occurs; an alarm message storage unit that stores an alarm message corresponding to the abnormal situation; and an alarm message storage unit that stores a warning message corresponding to the abnormal situation; An alarm message selection section that selects an alarm message, a speech synthesis section that performs speech synthesis based on the alarm message output of the alarm message selection section, and an audio output section that outputs speech in response to the output of the speech synthesis section. The audio alarm device has an input section for setting an alarm message, the speech synthesis section includes an LSI for speech synthesis using a waveform encoding method, and the alarm message storage section includes an LSI for alarm message storage.
A voice warning device comprising an AM, and further comprising a noise removal section for removing noise such as granular noise contained in the output of the voice synthesis section between the voice synthesis section and the voice output section. (2) The audio warning device according to claim 1, wherein the noise removal section includes a lower limiter circuit.