JPS58177294A - Emergency stop device for robot - 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 The present invention relates to an emergency stop device for a robot to ensure the safety of workers working around an industrial robot.

Conventionally, most of the methods for emergency stopping industrial robots in response to abnormal operation have been by using an emergency stop button provided on the robot's teaching box or control box. For this reason, workers always carry a teaching pendant to ensure safety from abnormal robot movements.

However, with this conventional emergency stop method, a worker who encounters an abnormal operation of the robot must operate the emergency stop button himself, which may result in cases where there is no time to operate the button or there is a delay in button operation. Assuming such situations, it is insufficient to ensure the safety of workers.

Therefore, it has been desired to realize an emergency stop device that can stop a robot using the voice emitted by a worker.

However, when configuring an abnormal stop device using voice recognition,
It is necessary to ensure that the system operates reliably in response to the emergency voice of a worker who is actually experiencing an accident, and to prevent it from inadvertently malfunctioning due to surrounding noise.

Therefore, the reliability of the device is particularly important, and from this point of view, conventionally it has been difficult to realize an emergency stop device using voice recognition.

This invention was made in view of the problem described in -E.The purpose of this invention is to be able to bring the robot to an emergency stop using the voice emitted by the worker, and to constantly perform self-diagnosis to check the operating status of the voice recognition unit. An object of the present invention is to provide an emergency stop device for a robot that can guarantee high reliability by checking the following.

Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

FIG. 1 is a schematic block diagram showing the electrical configuration of an emergency stop device for a robot according to the present invention.

In the same figure, this device is a microphone 1 that is fixedly installed around the robot installation position or movably installed on the clothes of the worker, and that collects the sounds emitted by the worker around the robot installation position. Then, limit the audio output of this microphone 1 (for example, from 1001-12 to 5KH2).
In response to the regular output command a issued by the control unit 4 based on the timing signal of the amplifier 2 and the timer 3,
A model audio generator 5 that emits a model audio signal having the same band as the above band, an adder 6 that adds each output signal of the amplifier 2 and the model audio generator 5 in an analog manner, and an adder 6
A plurality of bandpass filters 7Δ separate the audio signal of the addition output into a plurality of frequency bands.

7B,...7J (100Hz~ in this example)
5KH2 frequency band with 10 band pass filters 7
8 to 7J) and each filter 7A to 7
10 detectors 8A and 8B each detecting the output of J.
. . . 8J, ten level discriminators 9A, 9B . . . 9J, and level Create an interrupt signal using the discrimination output of any of the outputs of the discriminators 9A to 9J and the model audio output command a.
R circuit 10, ten A/D converters 11A, 11B, . voice recognition means for recognizing the emergency voice of the employee collected by the microphone 1 and the model voice generated by the model voice generator 5 based on each digital output of the voice recognition means; and emergency voice recognition by the voice recognition means. In response to the output, the control unit 4 outputs a switching command C to disconnect the robot power switch 12, and the stop control means stops the robot in an emergency. Whether the model voice recognition output of the voice recognition means is within a predetermined evaluation value. a self-diagnosis means for determining whether the abnormality is detected by the self-diagnosis means; a means for driving the abnormality alarm 13 when an abnormality is detected by the self-diagnosis means; and means for correcting an evaluation function in the self-diagnosis means according to the abnormality; ,
A CPU (Central Processing Unit) 14 that controls the control section 4 and the A/D converters 11A to 11J, receives an interrupt signal, and operates as each means such as the voice recognition means and self recognition means, and the CPtJ14. A system memory 15 in which system programs to be executed are stored, and a CPt
A working memory 16 is used as a working area when the J15-4 executes a system program. Note that the working memory 16 has a power supply 17 for power failure compensation.

2 and 3 are flowcharts showing the system program executed by the CPU 14, focusing on the parts related to the present invention. FIG. 2 shows the operation of the model voice generation means, and FIG. The operation of each means is shown respectively. The control operation of the CPIJ 14 will be explained below according to these flowcharts.

In FIG. 2, the first step 200 is to read the contents of timer 3, which generates a timing signal. In step 201, it is determined whether or not timer 3 has timed up. If the time is up, a timing signal is output to the control section 4 at step 202, and a model voice flag F is set in the working memory 16. The control unit 4 receives this timing signal and controls the model voice generator 5.
A model audio output command a is output to. Next, timer 3 is reset at step 6-tub 203, and the process returns to step 200. The above operations are repeated, and the model voice generator 5 periodically generates the model voice signal having the above-mentioned contents. As is clear from FIG. 1, this model voice signal passes through the circuit related to voice recognition means after the adder 6, and
An interrupt signal periodically interrupts the PU 14. Further, the output command a directly serves as an interrupt signal. CPU14
In response to this periodic interrupt signal, the self-diagnosis means, which will be described later, is periodically operated. As a result, an abnormality in the circuit related to the voice recognition means is quickly detected, and high reliability of the present device is ensured.

Next, in FIG. 3, when the CPU 14 receives an interrupt signal, it first checks the state of the model voice flag F in step 300. This is a routine for determining whether the cause of the interruption is an audio signal from the microphone 1 (hereinafter referred to as actual audio) or a model audio signal. If the model audio flag F is not set, the process advances to step 301, and each digital output of the A/D converters 11A to 11J is read. Next, in step 302, the voltage values Va, Vd, . The evaluation value Q for the actual audio signal at the time of occurrence of an interrupt is determined from the following equation given by:

Note that the initial evaluation value QO is obtained in advance from a sample of emergency sounds that can be collected by the microphone 1.

Q=Xa Va 10Xb Vb +-+Xj Vj
(1) Then, in the next step 303, the difference ΔQ=Q−QO between the above-mentioned push values Q and QO is obtained, and from this it is determined as follows whether or not the audio signal is an emergency audio signal.

Ra=i:△Q>O:W4m audio signal RO=O:ΔQ〈0: Normal audio signal Therefore, if Ra=1, the control unit 4 is driven in step 304 to output the opening/closing command C . With this, the robot power switch 12 is cut off, and the robot comes to an emergency stop. Further, if Ra = O in step 303, the signal is a normal audio signal, so step 304 is skipped.

Further, if the model voice flag F is set in step 300, the process advances to step 305. In step 305, each digital output of the A/D converters 11A to 11J is read, and the model audio flag F is reset. Next, in step 306, it is determined whether the actual audio signal is superimposed on the model audio signal. If an actual voice signal is included, the process proceeds to step 302, where the above-mentioned evaluation calculation is performed, and if the signal is an emergency voice, the robot is brought to an emergency stop. If there is only a model audio signal, the process proceeds to step 307 and subsequent steps to perform self-diagnosis. Step 307
Now, from each digital output obtained in step 305, the voltage values Va, Vb, etc. of each detection output of the detectors 8A to 8J are calculated.
... Find Vj, and use these and certain coefficients Ya, Yb...
...The evaluation value Q for the model audio signal at the time of interrupt occurrence is determined from the following equation given by Yj. Note that the initial evaluation value QO of the model audio signal is determined in advance.

Q=YaVa+YbVb・Yjl (2>9- Then, in the next step 30B, the above-mentioned pressing values Q and QO
The difference ΔQ=Q−QO is determined, and from these, it is determined as follows whether the evaluation value for the model audio signal at the time of occurrence of the interrupt is within the allowable range.

Rb=1: ΔQ>O: Evaluation l1lli value is outside the allowable range Rb
=O: ΔQ<O: Evaluation value is within the allowable range As a result, Rb
If =O, this device is normal, so step 30
9.Skip 310.

Moreover, if Rb=1, each routine of steps 309 and 310 is executed. Rb=1 means that adder 6
, bandpass filters 78-7J.

Detectors 8A to 8J, level discriminators 9A to 9J and A/
This means that an abnormality has occurred or is about to occur in any of the circuits related to the voice I recognition means such as the D converters 11A to 11J. Therefore, in this step 309, the abnormality alarm 13 is activated. By the way, even if the abnormality alarm 13 is activated, abnormality recovery measures are not necessarily taken immediately, and even within the period until the abnormality recovery measures are taken, the stop i
A facility may need to maintain its functionality.

-10- Therefore, in the next step 310, the evaluation function is corrected. The evaluation value Q determined to be Rb=1 in step 308 is set as Qa again, and ΔQa =Qa −Q.

5=Qa/Q.

Find N=1/S+α (α>O). Then, the coefficients of the modified equation (2) are Pa, p
b...Pj, the following relationship exists between the coefficients before and after correction.

Pa = N -Ya Pb = N-Yb pj = N-y, + Abnormality recovery action is taken using the vector P-(Pa, Pb -------Pj) modified in this way Perform a self-diagnosis during the period up to.

This allows the device to continue operating while retaining the ability to emergency stop the robot via voice in the event of a minor failure.

As explained in detail above, the emergency stop device for a robot according to the present invention transmits tl and h an emergency voice emitted by a human,
With this, the robots 1 to 1 can be brought to an emergency stop, so the worker does not need to carry a teaching box as in the past, and can work safely. Furthermore, since self-diagnosis is performed using periodic model voice signals, abnormalities in the voice recognition section can be quickly detected, and the reliability of the device is ensured. Furthermore, since it is possible to operate this device to its maximum potential by modifying the evaluation function, this device has the ability to make an emergency stop of the robot even before abnormality recovery measures are taken, ensuring the safety of workers. can.

[Brief explanation of the drawing]

FIG. 1 is a schematic block diagram showing an embodiment of the emergency stop device for robots 1 to 1 according to the present invention, FIGS. 2 and 3 are flowcharts of the control operation of the device for the emergency □ stop, and FIG. 3 shows the operation of the model voice generation means, and FIG. 3 shows the operation of each means such as the voice recognition means. 1...Microphone 5...Model audio generator 12...
・Robot power switch 13... Abnormality alarm 14... CPU 15... System memory 16... Working memory F...・Model sound flag Patent applicant Tateishi Electric Co., Ltd. 13- Figure 2 Figure 3

Claims (1)

[Claims] (1) A microphone that collects sounds emitted by humans around the robot installation position, and a voice recognition means that recognizes specific emergency sounds emitted by humans from the output of this microphone;
stop control means for stopping the robot in response to the emergency voice recognition output of the voice recognition means; and model voice generation means for periodically inputting a predetermined voice signal to the voice recognition means together with the output of the microphone. , self-diagnosis means for determining whether the recognition output of the voice recognition means for the model voice signal from the model voice generation means is within a predetermined evaluation value; and a self-diagnosis means for issuing an alarm when an abnormality is detected by the self-diagnosis means. An emergency stop device for a robot, comprising: means for generating a signal and correcting an evaluation function in the self-diagnosis means according to the abnormality.