Work status detection device, program, and work status detection method
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- PANASONIC INTELLECTUAL PROPERTY CORP OF AMERICA
- Filing Date
- 2021-12-17
- Publication Date
- 2026-08-03
Smart Images
Figure 0007899158000001 
Figure 0007899158000002 
Figure 0007899158000003
Abstract
Description
Technical Field
[0001] The present disclosure relates to a work status detection device, a program, and a work status detection method for detecting the progress of cleaning work or the like in an indoor space having a large number of seats, such as a train, an airplane, a ship, a theater, or a cinema hall. [[ID=⑥]]<0000⑩⑤>[[ID=⑦]]
Background Art
Prior Art Documents
Patent Documents
Patent Document 1
Summary of the Invention
Means for Solving the Problems
Brief Description of the Drawings
[0007] [Figure 1] This figure schematically shows the configuration of the work status detection device according to an embodiment of the present disclosure. [Figure 2] This is a schematic diagram showing the interior of a train car. [Figure 3] This is a schematic diagram showing the interior of a train car. [Figure 4] This diagram shows the process of changing pillowcases. [Figure 5] This figure shows an example of the frequency characteristics of work noise. [Figure 6] This figure shows another example of the frequency characteristics of work noise. [Figure 7] This diagram shows the internal structure of a noise control filter. [Modes for carrying out the invention]
[0008] (Knowledge that forms the basis of this disclosure) Normally, trains operate a round trip between their starting and ending stations, with passengers getting on and off at stations along the way, and finally all passengers disembarking at the final station. Then, the train reverses direction and operates from this final station to become the next starting station. During this time, especially on express trains, cleaning work is carried out mainly around the seats, such as returning reclined seats to their original position, rotating seats to change their direction, and changing pillow covers.
[0009] Similarly, on aircraft, all passengers who board at the departure point disembark at the arrival point, so the time between boardings is used to carry out cleaning tasks, mainly such as tidying up the seating area and changing pillowcases.
[0010] Furthermore, in theaters and cinemas, once a performance or screening has ended and the audience has left, cleaning work is carried out during the time before the next audience enters, mainly tidying up the seating area and returning seats to their original reclining positions.
[0011] Thus, completing cleaning tasks at multiple locations within a limited time requires a large number of cleaning workers to work simultaneously, and there is a need for increased efficiency and better progress management in this process.
[0012] Although it does not specifically target trains, airplanes, or movie theaters as described above, the technology disclosed in Patent Document 1 allows an information management center to track the location of each cleaning worker by having them carry an IC card with wireless functionality (transponder card), and then uses that information to issue work instructions to each worker, thereby improving overall work efficiency.
[0013] However, the technology disclosed in Patent Document 1 only allows for tracking which rooms have been cleaned by managing the current location of cleaning workers carrying transponding cards; it does not allow for detailed information such as which of the numerous cleaning items in each room is currently being performed. If the detailed current work could be known, appropriate preparations could be made for support, but without this information, such preparations cannot be made. In other words, it is not possible to further improve the efficiency of the entire cleaning operation. Furthermore, it is necessary to install a fixed station in each room to receive wireless information from the transponding card, and when applied to trains, airplanes, or movie theaters, where rooms are not separated, it is difficult to accurately determine the location of cleaning workers. In addition, it is necessary to install new fixed stations, which leads to challenges such as increased weight and increased costs.
[0014] To solve these problems, the inventors have come up with the idea of this disclosure based on the finding that it is possible to detect with high accuracy whether or not work has actually been performed by detecting the work sounds generated when a predetermined task is performed indoors and comparing these detected work sounds with representative work sounds that have been stored in advance.
[0015] Next, we will describe each aspect of this disclosure.
[0016] A working condition detection device according to an aspect of the present disclosure is installed indoors and includes at least one microphone that detects a first working sound, which is a working sound generated when a predetermined work is performed indoors, a storage device that stores in advance a second working sound, which is a representative working sound generated when the work is performed, a calculator that calculates the similarity between the first working sound and the second working sound, and a judge that determines that the work has been performed when the similarity is greater than or equal to a predetermined value.
[0017] According to this aspect, when the similarity between the first working sound detected by the microphone and the second working sound stored in the storage device in advance is greater than or equal to a predetermined value, the judge determines that the work has been performed. As a result, it is possible to accurately detect whether the work has actually been performed by comparing the working sounds.
[0018] In the above aspect, when a signal of the first working sound is input from the microphone and the signal level in a frequency band defined by a first threshold among the input signals of the first working sound is greater than or equal to a second threshold, a generated sound detector further extracts a signal component of the frequency band and inputs it to the calculator.
[0019] According to this aspect, the generated sound detector can appropriately detect the timing when a predetermined working sound is generated. As a result, the calculation accuracy of the similarity is improved, and it is possible to improve the detection accuracy of the predetermined work.
[0020] In the above aspect, when the ratio of the number of frequency samples whose signal level is greater than or equal to the second threshold to the total number of frequency samples included in the frequency band is greater than or equal to a third threshold, the generated sound detector extracts a signal component of the frequency band and inputs it to the calculator.
[0021] According to this aspect, the calculation accuracy of the similarity is further improved, and it is possible to further improve the detection accuracy of the predetermined work.
[0022] In the above embodiment, the work is performed at multiple locations within the room, the at least one microphone includes multiple microphones installed at the multiple locations, the determination device determines whether the work has been performed at each of the multiple locations, and further comprises a progress checker that manages the progress of the work at each location based on the determination result by the determination device, and an output device that outputs the management result by the progress checker and presents the information to the administrator.
[0023] According to this configuration, the manager can grasp the progress of work at each of the multiple locations within the room, thereby making it possible to improve the overall efficiency of the work.
[0024] In the above embodiment, a plurality of seats are installed in the room, the at least one microphone includes a plurality of microphones installed corresponding to each of the plurality of seats, the computer calculates the similarity between the first work sound and the second work sound with respect to the plurality of microphones corresponding to each seat, and the judge determines that the work has been performed on the seat if the similarity with respect to the plurality of microphones corresponding to each seat is equal to or greater than a predetermined value.
[0025] According to this embodiment, by using multiple microphones installed in relation to each seat, it becomes possible to further improve the detection accuracy of predetermined tasks.
[0026] In the above embodiment, a plurality of seats are installed in the room, and the at least one microphone includes a noise microphone and an error microphone installed corresponding to each of the plurality of seats, and further comprises a signal processing unit that generates a control signal by performing predetermined signal processing on a noise signal input from the noise microphone using a control coefficient updated based on an error signal input from the error microphone, and a speaker installed corresponding to each of the seats that outputs the control signal input from the signal processing unit.
[0027] According to this embodiment, the noise microphone and error microphone provided by the active noise control system can be repurposed as microphones for detecting work noises, thereby enabling miniaturization and cost reduction.
[0028] A program according to one aspect of the present disclosure is a program for causing a computer, which is a work status detection device, to function as a calculation means for calculating the similarity between the first work sound and the second work sound, and a determination means for determining that the work has been performed when the similarity is greater than or equal to a predetermined value. The computer comprises at least one microphone installed in a room for detecting a first work sound, which is a work sound generated when a predetermined work is performed in the room, and a memory in which a second work sound, which is a typical work sound generated when the work is performed, is pre-stored.
[0029] According to this embodiment, the determination means determines that work has been performed if the similarity between the first work sound detected by the microphone and the second work sound stored in the memory beforehand is greater than or equal to a predetermined value. As a result, it becomes possible to detect with high accuracy whether or not work has actually been performed by comparing the work sounds.
[0030] A work status detection method according to one aspect of the present disclosure includes a work status detection device comprising: at least one microphone installed in a room for detecting a first work sound, which is a work sound generated when a predetermined work is performed in the room; and a memory device in which a second work sound, which is a typical work sound generated when the work is performed, is stored in advance. The device calculates the similarity between the first work sound and the second work sound, and determines that the work has been performed if the similarity is equal to or greater than a predetermined value.
[0031] According to this embodiment, the work status detection device determines that work has been performed when the similarity between a first work sound detected by the microphone and a second work sound pre-stored in the memory is greater than or equal to a predetermined value. As a result, it becomes possible to detect with high accuracy whether or not work has actually been performed by comparing the work sounds.
[0032] This disclosure can also be implemented as a program that causes a computer to execute each characteristic configuration included in such a device, or as a system that operates using this program. It goes without saying that such a computer program can be distributed via a computer-readable, non-temporary recording medium such as a CD-ROM, or via a communication network such as the Internet.
[0033] (Embodiments of the present disclosure) Embodiments of this disclosure will be described in detail below with reference to the drawings. Elements denoted by the same reference numeral in different drawings refer to the same or corresponding elements.
[0034] The embodiments described below all represent preferred specific examples of the present disclosure.
[0035] Furthermore, in the embodiments described below, an example of a system in which a microphone for detecting work noise is installed is shown as active noise control (ANC), which reduces ambient noise for occupants seated in their seats, but the system is not limited to this.
[0036] Furthermore, the components, their arrangement and connection configurations, and the sequence of operations shown in the following embodiments are merely examples and are not intended to limit the present invention. The present invention is limited solely to the claims.
[0037] Therefore, among the components in the following embodiments, those components that are not described in the independent claim representing the highest-level concept of the present invention are described as constituting a more preferable form, even though they are not necessarily required to achieve the objectives of the present invention.
[0038] The configuration of the work status detection device according to the embodiment of this disclosure will now be described. Figure 1 is a schematic diagram showing the configuration of the work status detection device according to the embodiment of this disclosure. Figure 2 is a schematic diagram showing the interior of train 1000 as an example of application of this disclosure.
[0039] The work status detection device in Figure 1 is shown as being applied to the interior of train 1000, as shown in Figure 2. Figure 2 illustrates the seating arrangement in the interior of train 1000 as seen from above. As shown, the seats in train 1000 are typically configured as two or three seats joined together, such as seats 1D, 1E or seats 1A, 1B, 1C. These are arranged in a row with an aisle in between, and multiple such rows (10 rows in Figure 2) are installed. Figure 1 shows a case where two seats are joined together, such as seats 5D, 5E or seats 6D, 6E in Figure 2.
[0040] Figure 1 illustrates a top-down view of the interior of train 1000. Microphones 2a and 2b and speakers 3a and 3b are installed on the window-side seat 100a, and microphones 2c and 2d and speakers 3c and 3d are installed on the adjacent aisle-side seat 100b. Microphones 1a to 1h are also installed near the window 1101 on the vehicle panel 1100 on the side of seat 100a. Furthermore, microphones 1i and 1j are installed on the rear of the seatback of the seat 110a in the row in front of seat 100a, and microphones 1k and 1l are installed on the rear of the seatback of the seat 110b in the row in front of seat 100b.
[0041] Figure 3 is a schematic diagram of the interior of train 1000, representing a side view of the interior and showing seat 100a by the window. Although not shown in Figure 3, seat 100b is located in front of seat 100a. As shown in Figure 3, microphones 1a to 1f are installed on the vehicle panel 1100 near window 1101 so as to be symmetrically arranged. Microphones 1g and 1h are installed on the vehicle panel 1100 or luggage rack 1400 so as to be located near the luggage rack 1400 near window 1101. Although not shown in Figure 3, microphones similar to microphones 1a to 1h are also installed near window 1111. Furthermore, microphones 1i and 1j are installed on the rear side of the seat back of seat 110a.
[0042] Referring to Figure 1, the detection signals from microphones 1a to 1l are input to the noise control filter 500.
[0043] The detection operation for the replacement of pillow covers 101 (101a, 101b) placed on the headrests of seats 100a and 100b will be explained with reference to Figure 1. Figure 4 shows the process of replacing pillow covers 101. As shown in Figure 4, the replacement of pillow covers 101 is performed by a cleaning worker removing the used pillow cover 101 and putting on a new one. Since pillow covers 101 are usually secured with Velcro (registered trademark), a loud noise is generated when the pillow cover 101 is removed. Naturally, noises related to the work are generated throughout the series of replacement operations, including when attaching the pillow cover. The work status detection device according to this embodiment detects these noises and determines that the replacement of pillow covers 101 is being performed.
[0044] In Figure 1, for example, if the pillow cover 101b is to be replaced on seat 100b, the worker first removes the pillow cover 101b. This generates noise, which is detected by a microphone 2d near the pillow cover 101b. The signal of this noise is input from the microphone 2d to the similarity calculator 22d via the sound detector 21d.
[0045] The sound data memory 40 has pre-stored sound data related to typical work sounds generated during the replacement of the pillow cover 101. The similarity calculator 22d calculates the similarity between the work sound signal input from the microphone 2d via the sound detector 21d and the sound data input from the sound data memory 40. The signal indicating the calculation result by the similarity calculator 22d is input to the work presence / absence determination device 30b. The work presence / absence determination device 30b determines that "the replacement of the pillow cover 101b of seat 100b has been performed" if the similarity input from the similarity calculator 22d is equal to or greater than a predetermined value. The signal indicating the determination result by the work presence / absence determination device 30b is input to the work progress confirmation device 50.
[0046] Here, a pair of microphones 2c and 2d are installed on seat 100b, flanking the pillow cover 101b, which is the source of the work noise. Similar to microphone 2d, microphone 2c detects the work noise of the pillow cover 101b replacement work. Then, similar to the similarity calculator 22d, the similarity calculator 22c calculates the similarity between the work noise signal input from microphone 2c via the sound detector 21c and the sound data input from the sound data storage 40. The signal indicating the calculation result by the similarity calculator 22c is input to the work presence / absence determination 30b. The work presence / absence determination 30b determines that "the pillow cover 101b of seat 100b has been replaced" if the similarity values input from both similarity calculators 22c and 22d are equal to or greater than a predetermined value. On the other hand, the work presence / absence determination device 30b determines that "the work of replacing the pillow cover 101b of seat 100b has not been performed" if at least one of the similarity scores input from the similarity calculators 22c and 22d is less than the predetermined value. In this way, by using the sound signals from both microphone 2c and microphone 2d, the accuracy of the work presence / absence determination device 30b can be improved.
[0047] Here, we will explain how the similarity is calculated by the similarity calculator 22. Figure 5 shows an example of the frequency characteristics of a work sound. Assume that the work sound signals detected by microphones 2c and 2d have the frequency characteristics K1A shown in Figure 5. On the other hand, assume that the representative sound data stored in the sound data memory 40 has the frequency characteristics K2A shown in Figure 5. In this case, the simplest similarity calculation is to find the difference in level values for each frequency sample, find the number of samples that fall within a certain range, and determine the similarity based on that number. Specifically, if there are 8192 frequency samples (horizontal axis) each for the work sound signal (frequency characteristics K1A) and the representative sound data (frequency characteristics K2A), the similarity calculator 22 will find the difference in signal levels (vertical axis) of frequency characteristics K1A and K2A for each frequency sample, and if, for example, 6000 samples fall within a certain tolerance range (e.g., ±5dB), then the similarity will be set to 6000. The work presence / absence determination device 30 determines that the pillow cover 101 replacement work has been performed if the similarity calculated by the similarity calculator 22 is, for example, 5000 or more, while determining that the pillow cover 101 replacement work has not been performed if the similarity is, for example, less than 5000. Note that if the work sound exhibits characteristics more clearly in its time characteristics than in its frequency characteristics, the similarity may be calculated using the time characteristics instead of the frequency characteristics shown in Figure 5.
[0048] For a more mathematically precise method of determining similarity, one can calculate the vector distance between the working sound signal and the representative sound data. The shorter the vector distance, the higher the similarity. For example, if frequency characteristics like those shown in Figure 5 are obtained, cepstrum distance or similar methods can be used. If working sound signals and representative sound data representing time characteristics rather than frequency characteristics are used, one could, for example, calculate time variations, obtain feature vectors based on the calculated time variations, and then calculate the vector distance between the obtained feature vectors.
[0049] As described above, the similarity calculators 22c and 22d calculate the similarity between the work sound signal during the pillow cover 101b replacement work and the representative sound data from the sound data memory 40, and the work presence / absence determination device 30b determines that the pillow cover 101b replacement work has been performed if the similarity is equal to or greater than a predetermined value.
[0050] Here, the interior of train 1000 contains a mixture of various noises generated by multiple cleaning workers, as well as external noise and indoor air conditioning noise, which together constitute the background noise (background noise characteristic K3A) shown in Figure 5. Therefore, if the noise from the pillow cover replacement work is such that it is buried in the signal level of this background noise, the similarity calculator 22 will not be able to accurately calculate the similarity, and the work presence / absence judge 30 will not be able to accurately determine whether or not the pillow cover replacement work has been carried out. To address this issue, the similarity calculator 22 compares the noise detected by a microphone installed near the source of the target noise with the sound data. As described above, if the target noise is the noise from the pillow cover replacement work, the similarity calculator 22 compares the noise detected by microphone 2 installed near the pillow cover 101 with the sound data. From a different perspective, the work status detection device uses microphone 2, which is closest to the pillow cover 101 (the object of the work), among the multiple microphones used for ANC (Active Noise Cancellation) described later, to detect whether the work is complete or not. By using microphone 2, which is close to the pillow cover 101, the work sounds of the pillow cover 101 replacement work can be detected at a high signal level, so a work sound signal with a sufficient level difference from the ambient noise signal level can be detected, as shown in Figure 5. However, even if a work sound signal with a sufficient level difference from the ambient noise signal can be detected, other sounds are also generated, including conversations between workers, so it is difficult to accurately detect when a sound that is highly likely to be the work sound of the pillow cover 101 replacement work occurred. Furthermore, if the timing of the work occurrence cannot be detected to some extent, the similarity calculator 22 will not know which part of the sound constantly detected by microphone 2 should be extracted in order to calculate the similarity. If the extracted part is inappropriate, the characteristics of the work sound signal as shown in Figure 5 cannot be obtained, and the similarity calculation will not yield appropriate results.
[0051] To address this issue, as shown in Figure 5, a frequency band AR1 in which a sufficient level difference between the representative sound data and the background noise signal is ensured is defined by thresholds Th1L and Th1H. Threshold Th1L defines the lower frequency limit F1 of frequency band AR1, and threshold Th1H defines the upper frequency limit F2 of frequency band AR1. Furthermore, a signal level L1 that allows sufficient separation of the representative sound data signal level from the signal level of the background noise signal is defined by threshold Th2. In Figure 1, microphone 2 constantly detects sound, and the generated sound detector 21 sequentially analyzes the frequency characteristics of the sound signal. The generated sound detector 21 determines that the sound of the pillow cover replacement work has occurred when it observes that the frequency characteristic K1A of the analyzed sound signal has a signal level of or greater than threshold Th2 in the frequency band AR1, which is greater than or equal to threshold Th1L and less than or equal to threshold Th1H. The generated sound detector 21 then uses this frequency band AR1 as the evaluation range and inputs the work sound signal with frequency characteristic K1A within that evaluation range into the similarity calculator 22. The similarity calculator 22 calculates the similarity by comparing the working sound signal and the representative sound data with respect to the frequency band AR1.
[0052] As a result of this approach, it becomes possible to detect with a high degree of probability the timing at which the sound of the pillowcase 101 replacement work occurs, and the similarity calculation in the similarity calculator 22 can also obtain appropriate results.
[0053] Here, the sound detector 21 should determine that the sound of the pillowcase 101 replacement work has occurred if the ratio of frequency samples with a signal level of Th2 or higher to the total number of frequency samples within the frequency band AR1 of the detected sound signal thresholds Th1L to Th1H is equal to or greater than the threshold Th3. This improves the detection accuracy of the target work sound. The threshold Th3 can be set to, for example, 80%.
[0054] To accurately detect the occurrence of work sound signals and accurately calculate the similarity with representative sound data, the thresholds Th1L, Th1H, and Th2 in Figure 5 should be set appropriately. However, depending on the type of work, work sounds with characteristics different from those in Figure 5 may be generated. Figure 6 shows another example of the frequency characteristics of work sounds. The work sound signals detected by microphones 2c and 2d have the frequency characteristics K1B shown in Figure 6. Also, the representative sound data stored in the sound data memory 40 has the frequency characteristics K2B shown in Figure 6. For example, if work sounds like those in Figure 6 are generated, it may not be possible to accurately detect the occurrence of work sounds with a single set of thresholds Th1L and Th1H. Therefore, multiple sets of frequency thresholds Th1 may be set, such as thresholds Th11L and Th11H (frequency band AR2 of the lower frequency limit F3 and upper frequency limit F4) and thresholds Th12L and Th12L (frequency band AR3 of the lower frequency limit F5 and upper frequency limit F6). Furthermore, regarding the signal level threshold Th2, multiple thresholds Th2 may be set, such as threshold Th21 (signal level L2) and threshold Th22 (signal level L3). In other words, multiple sets of thresholds Th1 and multiple thresholds Th2 can be set according to the characteristics of the working sound. Of course, the same applies to threshold Th3.
[0055] Referring to Figure 1, when the work presence / absence determination device 30 determines that the pillow cover 101 replacement work has been carried out, a signal indicating the result of that determination is input from the work presence / absence determination device 30 to the work progress confirmation device 50. The work progress confirmation device 50 also receives the status of each replacement work for pillow covers 101 of seats other than seats 100a and 100b (seats 110a, 110b, etc. in the same car and seats in other cars), and the progress of the replacement work for pillow covers 101 of all seats in the interior of train 1000 is managed by the work progress confirmation device 50.
[0056] Furthermore, the work progress checker 50 manages not only the replacement of the pillow cover 101, but also all other work statuses, such as cleaning the table 112a installed on the seat, cleaning the luggage rack 1400, cleaning the seat surface with a tabletop broom, cleaning the floor of the aisle etc. with a vacuum cleaner or buffing machine, rotating the seat to align the seat direction, or returning the seat reclining angle to its initial state.
[0057] For example, in the cleaning operation of table 112a shown in Figure 3, microphones 1i and 1j placed near table 112a detect the sound produced when table 112a is unfolded for cleaning and the sound produced when table 112a is folded back up after cleaning. The similarity between the detected sound and representative sound data for the same operation stored in the sound data memory 40 is then calculated.
[0058] Furthermore, for example, in the cleaning of the luggage rack 1400 above the seat 100a shown in Figure 3, the sounds of the cleaning work on the luggage rack 1400 are detected by microphones 1g and 1h installed above the seat 100a. The similarity between the detected sounds and representative sound data of the same work stored in the sound data memory 40 is then calculated.
[0059] Similarly, for other tasks such as cleaning the seat surface, cleaning the floor, rotating the seat, or returning the seat recline angle to its initial position, the similarity between the work sound detected by a microphone placed near the source of the target work sound and the representative sound data for the same task stored in the sound data memory 40 is calculated. For tasks that do not necessarily target all seats in the vehicle, such as rotating the seat or returning the seat recline angle to its initial position, the following processing is possible. The work presence / absence determination device 30 determines, for example, that if no work sounds for these tasks are detected in the vehicle, these tasks are not being performed in that vehicle. On the other hand, if work sounds for these tasks are detected in even one location within the vehicle, it determines that these tasks are being performed in that vehicle.
[0060] The work progress checker 50 determines the status of each task based on the similarity of each task calculated in this manner, and manages the progress of all tasks within train 1000. The work progress checker 50 then inputs a signal indicating the progress of all tasks to the progress display 60, which displays the information on the progress display 60. This allows the work manager, who is in charge of the overall interior cleaning of train 1000, to grasp the progress of each task in each car at a glance. Therefore, if any task is stalled in any car, it becomes possible to appropriately instruct workers who are making progress to assist with the delayed tasks in order to complete the work within the limited time. As a result, the overall work efficiency of train 1000 can be improved. Note that the method of displaying information indicating the progress of all tasks is not limited to screen display on the progress display 60, but may also be audio output, etc.
[0061] The signal transmission from the work progress confirmation device 50 to the progress indicator 60 may be performed wirelessly or using a fixed network system established within the train 1000. If performed wirelessly, the progress indicator 60 may be a portable terminal held by the work manager. If a fixed network system is used, the progress indicator 60 may be a display device installed at the location where the work manager is located, or the portable terminal held by the work manager may be connected to the fixed network system via a hub.
[0062] Here, the work status detection device in Figure 1 can utilize the ANC system used to reduce running noise inside the train 1000's cabin, and the core of the signal processing in the ANC system is the noise control filter 500. This operation will now be explained.
[0063] In the ANC system, microphone 1 (1a-1l) is a noise microphone for detecting noise signals, and microphone 2 (2a-2d) is an error microphone for detecting error signals. The noise control filter 500 processes the noise signal detected by noise microphone 1 using the signal processor 501 so that the noise is reduced at the location where error microphone 2 is installed, and reproduces it as a control sound (control signal) from speaker 3 (3a-3d). Then, the driving noise and the control sound interfere at the location where error microphone 2 is installed, and the error microphone 2 detects the residual signal (error signal). Normally, the noise control filter 500 updates its control coefficient to minimize this error signal using adaptive signal processing. By repeatedly performing this process, the error signal is minimized, and a control coefficient that reduces driving noise is obtained.
[0064] Let's explain ANC in more detail. Figure 7 shows the internal configuration of the noise control filter 500 shown in Figure 1.
[0065] In Figure 7, the noise signal detected by the noise microphone 1 is processed with control coefficients in the signal processor 501, and the control signal output from the signal processor 501 is output from the speaker 3. At the same time, the noise signal detected by the noise microphone 1 is processed with coefficients in the transfer characteristic corrector 502.
[0066] Here, the transfer characteristic corrector 502 has the transfer characteristics from speaker 3 to error microphone 2 pre-set as coefficients.
[0067] The coefficients set in this manner and the noise signal detected by the noise microphone 1 are processed by the transfer characteristic corrector 502, and its output is input to the coefficient updater 503.
[0068] The coefficient updater 503 uses the output signal from the transfer characteristic corrector 502 and the error signal from the error microphone 2 to update the coefficients of the signal processor 501 to minimize the error signal using adaptive signal processing such as the least squares method or the learning identification method. By repeatedly performing this adaptive signal processing, the error signal is minimized, and the optimal control coefficient for reducing running noise is determined. As a result of determining the control coefficient of the signal processor 501, the running noise at the installation location of the error microphone 2 is minimized.
[0069] Thus, if an ANC system already exists that includes a noise microphone 1, an error microphone 2, a speaker 3, and a noise control filter 500, it is possible to miniaturize and reduce the cost of the system by reusing these existing noise microphones 1 and 2 as microphones for the work status detection device. Since the ANC system uses multiple microphones, it becomes possible to manage the progress of various cleaning tasks at different locations in the room.
[0070] Furthermore, not limited to ANC systems, but taking audio services as an example, if there are systems that utilize multiple microphones, such as sound field control systems that control a wide area, directional control systems, or wavefront control systems, these systems may be combined with work status detection devices.
[0071] Furthermore, although the above embodiment shows application to trains as an example, it is not limited to this and may also be applied to aircraft, ships, theaters, or cinemas. In the case of transportation methods that serve a large number of passengers, such as trains, aircraft, or ships, it is necessary to complete the cleaning of the interior within a limited time between when the current passengers disembark and when the next passengers board. The work status detection device according to this embodiment allows for accurate understanding of the progress of the cleaning work, making it possible to improve overall work efficiency by dispatching support to areas where work is delayed. Similarly, in spaces that provide simultaneous viewing services to a large number of people, such as theaters or cinemas, it is necessary to complete the cleaning of the interior within a limited time between when the current performance or screening ends and when the current audience leaves and when the audience for the next performance or screening enters. The work status detection device according to this embodiment allows for accurate understanding of the progress of the cleaning work, making it possible to improve overall work efficiency by dispatching support to areas where work is delayed. [Industrial applicability]
[0072] This disclosure is particularly useful for application to work status detection devices that detect the progress of cleaning work or other tasks in rooms with a large number of seats, such as trains, aircraft, ships, theaters, or cinemas. [Explanation of symbols]
[0073] 1 (1a~1l), 2 (2a~2d) Microphone 3 (3a~3d) Speakers 21 (21a~21d) Sound generation detector 22(22a~22d) Similarity calculator 30(30a,30b) Work presence / absence judgment device 40-sound data storage device 50 Work progress checker 60 Progress indicator 500 Noise Control Filter 501 Signal Processor
Claims
1. Multiple microphones are installed indoors to detect a first work sound, which is a work sound generated when a predetermined task is performed in the said indoors. A memory device in which a second work sound, which is a typical work sound generated when the aforementioned work is performed, is pre-stored, A computer that calculates the similarity between the first work sound and the second work sound detected by the microphone closest to the source of the first work sound among the plurality of microphones, A determination device that determines that the above operation has been performed when the similarity is equal to or greater than a predetermined value, A work status detection device equipped with the following features.
2. The work status detection device according to claim 1, further comprising a sound generation detector that receives a signal of the first work sound from the microphone, and when the signal level in a frequency band defined by a first threshold is equal to or greater than a second threshold of the input signal of the first work sound, extracts the signal component in that frequency band and inputs it to the computer.
3. The work status detection device according to claim 2, wherein the sound generation detector extracts the signal components of the frequency band and inputs them to the computer when the ratio of the number of frequency samples whose signal level is equal to or greater than the second threshold to the total number of frequency samples included in the frequency band is equal to or greater than the third threshold.
4. The aforementioned work was carried out at multiple locations within the room. The aforementioned multiple microphones are installed at the aforementioned multiple locations. The judgment device determines whether the work has been performed with respect to each of the multiple locations. A progress checker manages the progress of the work at each of the aforementioned locations based on the judgment result of the aforementioned judgment device, An output device that outputs the management results from the progress checker and presents the information to the administrator, A work status detection device according to any one of claims 1 to 3, further comprising:
5. The aforementioned room is equipped with multiple seats, The aforementioned multiple microphones are installed corresponding to each seat of the aforementioned multiple seats, The calculator calculates the similarity between the first work sound and the second work sound with respect to the plurality of microphones corresponding to each seat. The work status detection device according to any one of claims 1 to 4, wherein the determination device determines that the work has been performed on a seat when the similarity with respect to the plurality of microphones corresponding to each seat is equal to or greater than a predetermined value.
6. The aforementioned room is equipped with multiple seats, The plurality of microphones include noise microphones and error microphones installed in relation to each of the plurality of seats. A signal processing unit generates a control signal by performing predetermined signal processing on a noise signal input from the noise microphone using a control coefficient updated based on an error signal input from the error microphone. A speaker is installed corresponding to each of the aforementioned seats and outputs the control signal input from the signal processing unit, A work status detection device according to any one of claims 1 to 5, further comprising:
7. The work status detection device according to any one of claims 1 to 6, wherein the determination device determines that the work has been performed when the similarity detected by two or more microphones among the plurality of microphones that are closest to the source of the first work sound is greater than or equal to a predetermined value.
8. Multiple microphones are installed indoors to detect a first work sound, which is a work sound generated when a predetermined task is performed in the said indoors. A memory device in which a second work sound, which is a typical work sound generated when the aforementioned work is performed, is pre-stored, A computer as a work status detection device equipped with, A calculation means for calculating the similarity between the first work sound and the second work sound detected by the microphone closest to the source of the first work sound among the plurality of microphones, A determination means for determining that the operation has been performed when the similarity is equal to or greater than a predetermined value, A program designed to function as such.
9. Multiple microphones are installed indoors to detect a first work sound, which is a work sound generated when a predetermined task is performed in the said indoors. A memory device in which a second work sound, which is a typical work sound generated when the aforementioned work is performed, is pre-stored, A work status detection device equipped with, The similarity between the first work sound and the second work sound detected by the microphone closest to the source of the first work sound among the plurality of microphones is calculated. A work status detection method that determines that the work has been performed when the similarity is equal to or greater than a predetermined value.