Elevator status detection method, device, equipment and storage medium

Abstract

The present application provides an elevator state detection method, apparatus, device, and storage medium. The method includes: performing a frequency domain transformation on a recently acquired acceleration sequence to obtain a target spectrum, and converting the target spectrum into a power spectrum; then, when the power value corresponding to the power spectrum jumps from a first power zone to a second power zone, starting to record real-time acceleration values, and stopping recording until the real-time acceleration values ​​fall within a preset critical range, thereby obtaining a target acceleration sequence; comparing the target acceleration sequence with a set state template sequence to determine the vertical motion state of the elevator; finally, performing a difference calculation on the elements in the recently acquired air pressure sequence and a reference air pressure value to obtain an air pressure difference sequence; inputting the air pressure difference sequence into a trained neural network model to obtain a prediction result; and determining the elevator floor position based on the prediction result. This solution effectively improves the accuracy of elevator state detection and provides a reliable state data reference.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of computer technology, and in particular to an elevator status detection method, apparatus, device, and storage medium. Background Art

[0002] At present, in order to ensure the safety and stability of elevator control, elevator manufacturers usually do not release data on the elevator's operating status. In order to facilitate service robots to perform cross-floor operations or to facilitate the elevator operation and maintenance background to monitor the elevator's operating status, the existing method is to add sensors such as accelerometers and barometers to the elevator, and determine the elevator's operating status and elevator floor position based on the collected sensor data.

[0003] However, in related technologies, the elevator operating status is usually determined directly by the instantaneous numerical change of acceleration, or the elevator floor position is determined directly by the single air pressure difference. This is prone to abnormal fluctuations in the sensor data caused by external interference, or error accumulation problems, resulting in abnormal elevator status detection results. Summary of the Invention

[0004] The embodiments of the present application provide an elevator status detection method, apparatus, device, and storage medium, which solve the problem of abnormal elevator status confirmation results due to abnormal fluctuations or error accumulation in sensor data. By acquiring the latest acceleration sequence and the latest air pressure sequence in real time, accidental data errors are reduced, and abnormal data fluctuations are effectively dealt with. The elevator operating status is determined by the frequency domain power changes of the acceleration sequence and the state template sequence, and the floor position is predicted based on the air pressure difference through a neural network model, thereby effectively improving the accuracy of elevator status detection and providing a reliable status data reference.

[0005] In a first aspect, an embodiment of the present application provides an elevator status detection method, the method comprising:

[0006] Acquire a recent acceleration sequence of a first preset length and a recent air pressure sequence of a second preset length during elevator operation, wherein acceleration data corresponding to the recent acceleration sequence is collected in real time by an accelerometer installed in the elevator car, and air pressure data corresponding to the recent air pressure sequence is collected in real time by a barometer installed in the elevator car;

[0007] Performing frequency domain transformation on the most recent acceleration sequence to obtain a target spectrum, and converting the target spectrum into a power spectrum;

[0008] When the power value corresponding to the power spectrum jumps from the first power zone to the second power zone, recording of the real-time acceleration value is started, and recording is stopped until the real-time acceleration value is within a preset critical range, to obtain a target acceleration sequence, and the target acceleration sequence is compared with a set state template sequence to determine the vertical motion state of the elevator, wherein the first power zone corresponds to a power value range when the elevator is in a stationary state, and the second power zone corresponds to a power value range when the elevator is in a variable speed motion state;

[0009] The elements in the most recent air pressure sequence are calculated to differ from the reference air pressure value to obtain an air pressure difference sequence, the air pressure difference sequence is input into the trained neural network model to obtain a prediction result, and the elevator floor position is determined based on the prediction result.

[0010] Optionally, the state template sequence includes an ascending state sequence;

[0011] The step of comparing the target acceleration sequence with a set state template sequence to determine the vertical motion state of the elevator includes:

[0012] Calculating a first sequence distance value between the target acceleration sequence and the ascending state sequence;

[0013] When the first sequence distance value is within a preset distance range, determining that the vertical motion state of the elevator is an ascending state;

[0014] When the first sequence distance value exceeds the preset distance range, it is determined that the vertical movement state of the elevator is a descending state.

[0015] Optionally, the state template sequence includes a descending state sequence;

[0016] The step of comparing the target acceleration sequence with a set state template sequence to determine the vertical motion state of the elevator includes:

[0017] Calculating a second sequence distance value between the target acceleration sequence and the descending state sequence;

[0018] When the second sequence distance value is within the preset distance range, determining that the vertical motion state of the elevator is a descending state;

[0019] When the second sequence distance value exceeds the preset distance range, it is determined that the vertical motion state of the elevator is an ascending state.

[0020] Optionally, the training process of the neural network model includes:

[0021] The calibrated air pressure data corresponding to the preset floor set is input into the constructed neural network model, the error value is calculated based on the output prediction result and the set standard result, and the model parameters are updated based on the gradient descent method. It is iterated until the error value reaches the preset threshold range to obtain a trained neural network model.

[0022] Optionally, determining the elevator floor position based on the prediction result includes:

[0023] Inputting the air pressure difference sequence into the trained neural network model to obtain a confidence set corresponding to all floors;

[0024] The floor corresponding to the element in the confidence set whose confidence reaches the set threshold is updated as the elevator floor position.

[0025] Optionally, before calculating the difference between the elements in the most recent air pressure sequence and the reference air pressure value to obtain the air pressure difference sequence, the method further includes:

[0026] Filtering the most recent air pressure sequence to obtain a filtered air pressure sequence;

[0027] Accordingly, the step of performing difference calculation between the elements in the most recent air pressure sequence and the reference air pressure value to obtain an air pressure difference sequence includes:

[0028] The elements in the filtered air pressure sequence are calculated to have differences with the reference air pressure value to obtain an air pressure difference sequence.

[0029] Optionally, after determining the vertical motion state of the elevator, the method further includes:

[0030] The vertical motion state is sent to the system background so that the system background updates the lifting state display of the elevator.

[0031] Optionally, after determining the elevator floor position based on the prediction result, the method further includes:

[0032] The elevator floor position is sent to the system backend so that the system backend updates the floor position display of the elevator.

[0033] Optionally, the elevator status detection method further includes:

[0034] In the case where the elevator floor position is different from the elevator floor position determined at a previous moment, adding the elevator floor position to a recorded floor change list;

[0035] The floor change list is checked based on the set inspection rules. If the inspection result is abnormal, the floor change list is saved to the log, and the reference floor is added to the queue of floors to be executed of the elevator to complete the update of the reference air pressure value.

[0036] Optionally, the elevator status detection method further includes:

[0037] When the elevator floor position is a reference floor, air pressure data of a preset time period is obtained, and a median value of the air pressure data is recorded as a reference air pressure value.

[0038] In a second aspect, an embodiment of the present application further provides an elevator state detection device, comprising:

[0039] an acquisition module configured to acquire a recent acceleration sequence of a first preset length and a recent air pressure sequence of a second preset length during elevator operation, wherein acceleration data corresponding to the recent acceleration sequence is acquired in real time by an accelerometer installed in the elevator car, and air pressure data corresponding to the recent air pressure sequence is acquired in real time by a barometer installed in the elevator car;

[0040] a frequency domain power conversion module configured to perform frequency domain transformation on the most recent acceleration sequence to obtain a target spectrum, and convert the target spectrum into a power spectrum;

[0041] a motion state determination module configured to, when the power value corresponding to the power spectrum jumps from a first power zone to a second power zone, start recording real-time acceleration values, stop recording until the real-time acceleration value is within a preset critical range, obtain a target acceleration sequence, compare the target acceleration sequence with a set state template sequence, and determine the vertical motion state of the elevator, wherein the first power zone corresponds to a power value range when the elevator is in a stationary state, and the second power zone corresponds to a power value range when the elevator is in a variable speed motion state;

[0042] The floor position determination module is configured to calculate the difference between the elements in the most recent air pressure sequence and the reference air pressure value to obtain an air pressure difference sequence, input the air pressure difference sequence into the trained neural network model to obtain a prediction result, and determine the elevator floor position based on the prediction result.

[0043] Optionally, the state template sequence includes an ascending state sequence; and the motion state determination module is configured to:

[0044] Calculating a first sequence distance value between the target acceleration sequence and the ascending state sequence;

[0045] When the first sequence distance value is within a preset distance range, determining that the vertical motion state of the elevator is an ascending state;

[0046] When the first sequence distance value exceeds the preset distance range, it is determined that the vertical movement state of the elevator is a descending state.

[0047] Optionally, the state template sequence includes a descending state sequence; the motion state determination module is configured to:

[0048] Calculating a second sequence distance value between the target acceleration sequence and the descending state sequence;

[0049] When the second sequence distance value is within the preset distance range, determining that the vertical motion state of the elevator is a descending state;

[0050] When the second sequence distance value exceeds the preset distance range, it is determined that the vertical motion state of the elevator is an ascending state.

[0051] Optionally, the training process of the neural network model includes:

[0052] The calibrated air pressure data corresponding to the preset floor set is input into the constructed neural network model, the error value is calculated based on the output prediction result and the set standard result, and the model parameters are updated based on the gradient descent method. It is iterated until the error value reaches the preset threshold range to obtain a trained neural network model.

[0053] Optionally, the floor location determination module is configured as follows:

[0054] Inputting the air pressure difference sequence into the trained neural network model to obtain a confidence set corresponding to all floors;

[0055] The floor corresponding to the element in the confidence set whose confidence reaches the set threshold is updated as the elevator floor position.

[0056] Optionally, a sequence filtering module is also included, configured as follows:

[0057] Filtering the most recent air pressure sequence to obtain a filtered air pressure sequence;

[0058] Accordingly, the motion state determination module is configured as follows:

[0059] The elements in the filtered air pressure sequence are calculated to have differences with the reference air pressure value to obtain an air pressure difference sequence.

[0060] Optionally, the system further includes a first sending module configured to:

[0061] The vertical motion state is sent to the system background so that the system background updates the lifting state display of the elevator.

[0062] Optionally, a second sending module is further included, configured as follows:

[0063] After determining the elevator floor position based on the prediction result, the method further includes:

[0064] The elevator floor position is sent to the system backend so that the system backend updates the floor position display of the elevator.

[0065] Optionally, a floor error correction module is also included, configured as follows:

[0066] In the case where the elevator floor position is different from the elevator floor position determined at a previous moment, adding the elevator floor position to a recorded floor change list;

[0067] The floor change list is checked based on the set inspection rules. If the inspection result is abnormal, the floor change list is saved to the log, and the reference floor is added to the queue of floors to be executed of the elevator to complete the update of the reference air pressure value.

[0068] Optionally, a reference air pressure module is also included, configured as follows:

[0069] When the elevator floor position is a reference floor, air pressure data of a preset time period is obtained, and a median value of the air pressure data is recorded as a reference air pressure value.

[0070] In a third aspect, an embodiment of the present application further provides an electronic device, the device comprising:

[0071] one or more processors;

[0072] a storage device configured to store one or more programs,

[0073] When the one or more programs are executed by the one or more processors, the one or more processors implement the elevator state detection method described in the embodiment of the present application.

[0074] In a fourth aspect, an embodiment of the present application further provides a non-volatile storage medium storing computer-executable instructions, wherein the computer-executable instructions are configured to execute the elevator state detection method described in the embodiment of the present application when executed by a computer processor.

[0075] In an embodiment of the present application, a recent acceleration sequence of a first preset length and a recent air pressure sequence of a second preset length are obtained during elevator operation. The recent acceleration sequence is then transformed in the frequency domain to obtain a target spectrum, which is then converted into a power spectrum. When the power value corresponding to the power spectrum jumps from a first power range to a second power range, recording of real-time acceleration values ​​is started and stopped until the real-time acceleration value falls within a preset critical range, thereby obtaining a target acceleration sequence. The target acceleration sequence is then compared with a set state template sequence to determine the vertical motion state of the elevator. Finally, the difference between the elements in the recent air pressure sequence and the reference air pressure value is calculated to obtain an air pressure difference sequence. The air pressure difference sequence is input into a trained neural network model to obtain a prediction result, and the elevator floor position is determined based on the prediction result. By acquiring the recent acceleration sequence and the recent air pressure sequence in real time, accidental data errors are reduced, abnormal data fluctuations are effectively addressed, the elevator operation state is determined by frequency domain power changes in the acceleration sequence and the state template sequence, and the floor position is predicted by the neural network model based on the air pressure difference, thereby effectively improving the accuracy of elevator state detection and providing a reliable state data reference. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 A flowchart of an elevator status detection method provided in an embodiment of the present application;

[0077] Figure 2 A flowchart of a method for determining the vertical motion state of an elevator provided in an embodiment of the present application;

[0078] Figure 3 A schematic diagram matching an example rising state sequence provided in an embodiment of the present application;

[0079] Figure 4 A schematic diagram of an example provided in an embodiment of the present application that does not match the rising state sequence;

[0080] Figure 5 A flowchart of another method for determining the vertical motion state of an elevator provided in an embodiment of the present application;

[0081] Figure 6 A schematic diagram matching a descending state sequence provided as an example in an embodiment of the present application;

[0082] Figure 7 A flowchart of another elevator status detection method provided in an embodiment of the present application;

[0083] Figure 8 A structural block diagram of an elevator status detection device provided in an embodiment of the present application;

[0084] Figure 9A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0085] The following is a further detailed description of the embodiments of the present application in conjunction with the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely for the purpose of explaining the embodiments of the present application, and are not intended to limit the embodiments of the present application. It should also be noted that, for ease of description, the accompanying drawings only illustrate portions of the embodiments of the present application, rather than all structures.

[0086] The terms "first," "second," and the like in the specification and claims of this application are used to distinguish similar objects, and are not used to describe a specific order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments of this application can be implemented in an order other than that illustrated or described herein, and that the objects distinguished by "first," "second," and the like are generally of the same type, and do not limit the number of objects; for example, the first object can be one or more. In addition, the term "and / or" in the specification and claims refers to at least one of the connected objects, and the character " / " generally indicates that the objects connected are in an "or" relationship.

[0087] The elevator status detection method provided in the embodiment of the present application can be used to provide elevator status data reference for service robots operating across floors, and can also provide data reference for monitoring the elevator operation status for the elevator operation and maintenance background, wherein the elevator status data includes the elevator operation status, such as the stationary state, the rising state and the descending state, as well as the elevator floor position. The several application scenarios listed above are only exemplary and explanatory. In actual applications, the elevator status detection method can also be used under the data requirements of other scenarios, and the embodiments of the present application do not limit this. The present application aims to provide an elevator status detection method to solve the problem of abnormal confirmation results of the elevator status due to abnormal fluctuations in sensor data or error accumulation.

[0088] In the elevator status detection method provided in the embodiment of the present application, the execution subject of each step can be a computer device, which refers to any electronic device with data calculation, processing and storage capabilities, such as mobile phones, PCs (Personal Computers), tablet computers and other terminal devices, or servers and other devices, or service robots, control terminals, etc., which are not limited in the embodiment of the present application.

[0089] Figure 1 This is a flow chart of an elevator status detection method provided in an embodiment of the present application. The elevator status detection method can be implemented with an elevator status detection device as the execution subject. Figure 1 As shown, the elevator status detection method specifically includes the following steps:

[0090] Step S101: Acquire a recent acceleration sequence of a first preset length and a recent air pressure sequence of a second preset length during elevator operation. The acceleration data corresponding to the recent acceleration sequence is collected in real time by an accelerometer installed in the elevator car, and the air pressure data corresponding to the recent air pressure sequence is collected in real time by a barometer installed in the elevator car.

[0091] An accelerometer and a barometer can be installed on the top of the elevator car. The accelerometer can be a single-axis accelerometer or a multi-axis accelerometer. For example, a three-axis accelerometer can output acceleration values ​​in the X and Y directions of the horizontal plane in real time, and can also output acceleration values ​​in the Z direction of the vertical plane in real time. Since elevators usually operate in the vertical direction, the Z direction data can be used as source data for elevator status detection. The barometer can output air pressure values ​​in real time when the elevator car is at different heights as source data for elevator floor position detection. In the actual application of accelerometers and barometers, the acceleration and air pressure values ​​collected are prone to varying degrees of noise due to external factors. As a result, the acceleration and air pressure values ​​will exhibit complex fluctuation characteristics. The instantaneous acceleration and air pressure values ​​cannot effectively represent the activity state and position state of the elevator. Relying solely on instantaneous values ​​for elevator status detection is prone to false detection, making the elevator status data unreliable. Therefore, the most recent acceleration sequence can be acquired in real time by using the first preset length as the reference data length. This is equivalent to acquiring continuous acceleration data or air pressure data by setting a fixed-length data window during the accelerometer data update process, while keeping the data window moving to acquire the latest acceleration data or air pressure data, thereby maintaining the timeliness of the data. For example, the first preset length can be set to 1024, so that the most recent acceleration sequence includes 1024 data, and the data is arranged in chronological order. Whenever the accelerometer updates the latest data according to the set sampling frequency, for example, the sampling frequency is 800Hz, the most recent acceleration sequence is equivalent to retaining 1.28 seconds of acceleration data, and continuously adding the latest acceleration data to the most recent acceleration sequence, while removing the oldest acceleration data, maintaining a fixed length of 1024 for the most recent acceleration sequence. For another example, the second preset length can be set to 10, then the most recent air pressure sequence includes 10 data, and the data is arranged in chronological order. Whenever the barometer updates the latest data according to the set sampling frequency, for example, the sampling frequency is 16Hz, then the most recent air pressure sequence is equivalent to retaining 0.625 seconds of air pressure data, and continuously adding the latest air pressure data to the most recent air pressure sequence, while removing the earliest air pressure data, maintaining a fixed length of 10 for the most recent air pressure sequence.

[0092] Therefore, by setting a preset length to obtain the most recent acceleration sequence and the most recent air pressure sequence, while retaining quantitative historical data and the latest data as a reference, the accidental data error is effectively reduced, the adverse effects of data noise are dealt with, and the accuracy of the detection results is improved.

[0093] Step S102: Perform frequency domain transformation on the most recent acceleration sequence to obtain a target spectrum, and convert the target spectrum into a power spectrum.

[0094] It is worth noting that since the recent acceleration sequence is time-domain data, the data fluctuates greatly due to noise, making it impossible to determine the elevator status by setting a threshold. Considering that the power intensity in the frequency domain corresponding to the stationary state, acceleration / deceleration state, and uniform speed state during elevator operation presents a step-by-step change, the stationary state, acceleration / deceleration state, and uniform speed state can be divided into corresponding power zones based on the power value distribution range in different states. Therefore, the recent acceleration sequence is transformed into the frequency domain to obtain the target spectrum, which is then converted into a power spectrum. The goal is to effectively use acceleration data from a power perspective to distinguish different operating states of the elevator, minimize the impact of noise data, and provide a reliable data foundation for subsequent detection and judgment.

[0095] Specifically, if the acceleration sequence collected during the operation of the elevator is directly subjected to Fourier transform to obtain frequency domain information, then the frequency domain information cannot reflect the characteristics of the acceleration changing with the time axis in the frequency domain, and lacks time characteristics. Therefore, performing spectrum transformation on the most recent acceleration sequence through short-time Fourier transform is equivalent to using a real-time moving data window to perform local spectrum transformation on the acceleration data sequence, and further solving to obtain the local power spectrum, and then obtaining a power spectrum diagram with time information. The signal power change characteristics are statistically obtained from the power spectrum, and the operating status of the elevator is identified and judged by the power area divided based on different power intensity thresholds. Optionally, taking the fast Fourier transform algorithm and window function as an example to perform local spectrum transformation, setting the first preset length to 1024, then the fixed length of the window function is 1024. By using the Hann window function, the problem of local spectrum leakage caused by phase discontinuity due to non-periodic truncation of the signal can be reduced. The specific discrete transformation calculation formula is as follows:

[0096]

[0097] Where w(k) is the window function, a(x) is the most recent acceleration sequence, h is the frame shift length, p is the number of frame shifts, and j is the imaginary unit.

[0098] Then, the frequency corresponding to the most recent acceleration sequence is further transformed to obtain the target spectrum, and the target spectrum is converted into a power spectrum to obtain a continuous power intensity distribution.

[0099] Step S103: When the power value corresponding to the power spectrum jumps from the first power zone to the second power zone, start recording the real-time acceleration value until the real-time acceleration value is within the preset critical range and stop recording to obtain a target acceleration sequence. The target acceleration sequence is compared with the set state template sequence to determine the vertical motion state of the elevator. The first power zone corresponds to the power value range when the elevator is in a stationary state, and the second power zone corresponds to the power value range when the elevator is in a variable speed motion state.

[0100] Among them, since the power threshold ranges corresponding to different operating states of the elevator are different, for example, when the elevator is in a stationary state, the change characteristics of the effective signal of the acceleration value tend to be gentle and the noise is small; when the elevator is in a uniform speed state, since the elevator is in a moving state, the acceleration value is compared with the stationary state, and the effective signal and noise are mixed, the amplitude and period are larger, and the overall change characteristics fluctuate significantly; when the elevator is in the acceleration and deceleration stage, the effective signal of the acceleration value changes with the speed change of the vertical movement of the elevator, and the amplitude is larger and the period is longer. Therefore, it can be regarded as that the stationary state, uniform speed state and acceleration and deceleration state correspond to power zones of different magnitudes, respectively, and this is used to distinguish the stationary state and moving state of the elevator.

[0101] Furthermore, the first power zone is set to correspond to the power value range when the elevator is in a stationary state, and the second power zone is set to correspond to the power value range when the elevator is in a variable speed motion state. Then, when the power value corresponding to the power spectrum jumps from the first power zone to the second power zone, it means that the elevator is changing from a stationary state to an accelerated motion state. The acceleration data can be cached and recorded to obtain a target acceleration sequence, and the target acceleration sequence can be compared with the set state template sequence to determine the vertical motion state of the elevator, where the state template sequence can be a reference data sequence set based on the law of acceleration change during the elevator motion. It is understandable that the upward or downward movement of an elevator can be subdivided into an acceleration phase, a uniform speed phase, and a deceleration phase. The acceleration change corresponding to the upward movement is to start from the acceleration reference value and increase positively to the peak value, then decrease to the acceleration reference value, then maintain a certain period of fluctuation around the acceleration reference value, then increase in the opposite direction to the peak value, and then decrease to the acceleration reference value. The acceleration change corresponding to the downward movement is to start from the acceleration reference value and increase in the opposite direction to the peak value, then decrease to the acceleration reference value, then maintain a certain period of acceleration reference value, then increase positively to the peak value, and then decrease to the acceleration reference value. The acceleration reference value can be set as the gravity acceleration value for the original collected target acceleration sequence. Since the acceleration variation patterns of an elevator during its upward or downward motion are exactly opposite, and the acceleration variation patterns of each upward or downward motion of the elevator are consistent, it is possible to set a state template sequence to match and confirm the different vertical motion states of the elevator. For example, with the positive direction of the time axis as a reference, a state template sequence can be established based on the overweight stage of the upward motion or only the acceleration increasing stage, or based on the weightless stage of the downward motion or only the acceleration decreasing stage. The target acceleration sequence is obtained by recording the real-time acceleration values ​​corresponding to the set state template sequence, and the target acceleration sequence is compared with the set state template sequence to determine the vertical motion state of the elevator. If the state template sequence is determined based on the overweight stage of the upward motion or the weightless stage of the downward motion, the preset critical range can be a range with a preset error floating up and down with the acceleration reference value as the center value, for example, with the gravity acceleration as the center value, and a preset error floating up and down, such as 0.2m / s. 2Etc., after recording begins, if the real-time acceleration value falls within a preset critical range after a period of time, it can be considered that the elevator has completed the overweight stage of the upward movement or the weightless stage of the downward movement. Therefore, the target acceleration sequence obtained can be compared with the corresponding state template sequence to determine whether the elevator is in an upward movement or a downward movement. If the state template sequence is determined based on the acceleration increasing stage of the upward movement or the acceleration decreasing stage of the downward movement, then the preset critical range can be a range with a preset error floating around the maximum value of the acceleration increasing stage as the center value, or a range with a preset error floating around the minimum value of the acceleration decreasing stage as the center value. When recording begins, if the real-time acceleration value falls within the preset critical range after a period of time, it can be considered that the elevator has completed the acceleration increasing stage of the upward movement or the acceleration decreasing stage of the downward movement. Therefore, the target acceleration sequence obtained can be compared with the corresponding state template sequence to determine whether the elevator is in an upward movement or a downward movement. From the perspective of reducing possible noise impact and improving accuracy, the former method can be used to refer to the overweight stage of the ascending movement or the weightlessness stage of the descending movement. From the perspective of timeliness, the latter method can be used to refer to the acceleration increasing stage of the ascending movement or the acceleration decreasing stage of the descending movement.

[0102] In one embodiment, the state template sequence includes a rising state sequence, Figure 2 A flowchart of a method for determining the vertical motion state of an elevator provided in an embodiment of the present application, based on the above embodiment, as shown in FIG. Figure 2 As shown, the specific implementation process of determining the vertical motion state of the elevator includes the following steps:

[0103] Step S1031: Calculate a first sequence distance value between the target acceleration sequence and the ascending state sequence.

[0104] Among them, the rising state sequence can be set based on the acceleration data characteristics of the overweight stage of the elevator's rising motion, or it can be set based on the data characteristics of the acceleration increasing stage of the elevator's rising motion. The first sequence distance value can be a measure of the similarity between the target acceleration sequence and the rising state sequence. The smaller the distance value difference, the higher the similarity, and the larger the distance value difference, the lower the similarity. Specifically, the calculation of the first sequence distance value can be completed by using the DTW (Dynamic Time Warping) algorithm, the FastDTW algorithm, etc., wherein, since the target acceleration sequence and the rising state sequence may be two groups of sequences of non-equal length, the DTW algorithm and the FastDTW algorithm can effectively perform similarity matching of sequences of non-equal length, avoiding the accuracy of the calculation results affected by inconsistent sequence lengths. Then, by judging whether the first sequence distance value is within a preset distance range, it is determined whether the target acceleration sequence matches the rising state sequence, and further determines whether the elevator is in an ascending state or a descending state.

[0105] Step S1032: When the first sequence distance value is within a preset distance range, determine that the vertical motion state of the elevator is an ascending state.

[0106] Step S1033: When the first sequence distance value exceeds the preset distance range, determine that the vertical motion state of the elevator is a descending state.

[0107] In the specific implementation process, Figure 3 The schematic diagram of the example provided in the embodiment of the present application that matches the rising state sequence is as follows: Figure 3 As shown, taking the DTW algorithm as an example, an ascending state sequence is set based on the acceleration data characteristics of the overweight stage of the elevator's ascending motion, and a first sequence distance value between the target acceleration sequence 301 and the ascending state sequence 302 is calculated. The first sequence distance value is within a preset distance range, indicating that the target acceleration sequence satisfies the process from increasing to decreasing, and it can be determined that the elevator is in an ascending state. Figure 4 The schematic diagram of the example provided in the embodiment of the present application does not match the rising state sequence, such as Figure 4 As shown, taking the DTW algorithm as an example, the first sequence distance value between the target acceleration sequence 303 and the ascending state sequence 304 is calculated. The first sequence distance value is not within the preset distance range, indicating that the target acceleration sequence does not satisfy the process from increasing to decreasing, and it can be determined that the elevator is in a descending state.

[0108] In one embodiment, the state template sequence includes a descending state sequence, Figure 5A flowchart of another method for determining the vertical motion state of an elevator provided in an embodiment of the present application. Based on the above embodiment, the specific implementation process of determining the vertical motion state of the elevator includes the following steps, wherein steps S1034 to S1036 can be replaced with steps S1031 to S1033 and have the same function:

[0109] Step S1034: Calculate a second sequence distance value between the target acceleration sequence and the descending state sequence.

[0110] The descending state sequence can be set based on the acceleration data characteristics of the weightlessness phase of the elevator's descending motion, or it can be set based on the data characteristics of the acceleration decreasing phase of the elevator's descending motion. The second sequence distance value can be a measure of the similarity between the target acceleration sequence and the descending state sequence. The smaller the distance value difference, the higher the similarity, and the larger the distance value difference, the lower the similarity. Specifically, the second sequence distance value can be calculated using a DTW (Dynamic Time Warping) algorithm, a FastDTW algorithm, or the like. Since the target acceleration sequence and the descending state sequence may be two sets of sequences of unequal length, the DTW algorithm or the FastDTW algorithm can effectively perform similarity matching of sequences of unequal length, avoiding the accuracy of the calculation results affected by inconsistent sequence lengths. Then, by determining whether the second sequence distance value is within a preset distance range, it is determined whether the target acceleration sequence matches the descending state sequence, and further determines whether the elevator is in a descending state or an ascending state.

[0111] Step S1035: When the second sequence distance value is within the preset distance range, determine that the vertical motion state of the elevator is a descending state.

[0112] Step S1036: When the second sequence distance value exceeds the preset distance range, determine that the vertical motion state of the elevator is the ascending state.

[0113] In the specific implementation process, Figure 6 The schematic diagram of the example provided in the embodiment of the present application and the matching of the descending state sequence is as follows: Figure 6 As shown, taking the DTW algorithm as an example, a descent state sequence is set based on the weightlessness stage of the elevator's descending motion, and a second sequence distance value between the target acceleration sequence 305 and the descent state sequence 306 is calculated. The second sequence distance value is within a preset distance range, indicating that the target acceleration sequence satisfies the process from decreasing to increasing, and it can be determined that the elevator is in a descending state.

[0114] Therefore, by distinguishing the elevator operating status from the perspective of the power distribution area, the impact of noise on distinguishing the elevator operating status is reduced, and then the set state template sequence is compared with the recorded target acceleration sequence to improve the accuracy of determining the vertical motion state of the elevator and ensure the timeliness of the detection results.

[0115] In one embodiment, based on the above embodiment, after determining the vertical motion state of the elevator, the following process is further included:

[0116] The vertical motion status is sent to the system background so that the system background can update the elevator's lifting status display.

[0117] It is worth noting that after the system background receives the vertical motion status, it can inform passengers at the elevator end whether the elevator is in an ascending or descending state by means of an indicator light, so that passengers can choose the appropriate elevator to wait according to the display of the indicator light. The service robot can also obtain the elevator's ascending and descending information based on the vertical motion status to further plan the path and synchronize the vertical motion status to the system background.

[0118] Step S104: Calculate the difference between the elements in the most recent air pressure sequence and the reference air pressure value to obtain an air pressure difference sequence, input the air pressure difference sequence into the trained neural network model to obtain a prediction result, and determine the elevator floor position based on the prediction result.

[0119] Due to the effects of cold and heat, atmospheric pressure fluctuates at different times, even on the same floor. This can affect the accuracy of the system's floor detection. Therefore, directly using absolute atmospheric pressure values ​​for floor detection can result in significant errors. Different elevator floor positions can be distinguished by using a relative reference to a reference atmospheric pressure value. Specifically, the reference atmospheric pressure value can be the atmospheric pressure value of the first floor or the atmospheric pressure value of a preset floor position. Taking the atmospheric pressure value of the first floor as an example, a Hall effect sensor can be installed on the top of the elevator car, and a guide tape can be installed on the first floor. When the Hall effect sensor follows the elevator's ascent and descent to reach the first floor, it senses the guide tape and records the measured atmospheric pressure value as the reference atmospheric pressure value. Thus, each time the elevator reaches the first floor, the reference atmospheric pressure value is re-determined, effectively adapting to atmospheric pressure fluctuations at different time periods and improving the accuracy of the reference data. Optionally, the specific process of determining the reference air pressure value can be that when the elevator car reaches the first floor, after the Hall sensor is in the triggered state, all air pressure values ​​within a preset time period can be recorded, and the median value can be calculated from all air pressure values, and the median value can be used as the reference air pressure value.

[0120] After determining the baseline air pressure value, a single air pressure value has fewer features and is easily affected by noise fluctuations. Therefore, a recent air pressure sequence of a second preset length is obtained. This approach improves interference resistance by referencing the changing characteristics of multiple consecutive air pressure values. The elements of the recent air pressure sequence are subtracted from the baseline air pressure value to generate an air pressure output value sequence. This pressure difference sequence is then input into the trained neural network model to generate a prediction result, and the elevator floor position is determined based on the prediction result.

[0121] In addition, due to the influence of external environmental conditions and the airflow disturbances caused by the elevator's ascent and descent, the barometer's raw real-time output is often superimposed with high-frequency noise as it follows the elevator car's ascent and descent. The actual change in pressure value typically has a signal frequency less than 20Hz in the spectrum, while the noise signal's frequency characteristic is typically greater than 20Hz. To restore the true change curve representing the elevator's operating status from the aliased, noisy pressure value signal, the most recent pressure sequence can be filtered to obtain a filtered pressure sequence. The difference between the elements in the filtered pressure sequence and the reference pressure value is then calculated to obtain a pressure difference sequence. For example, an IIR digital filter, an FIR digital filter, or the like can be used to remove the high-frequency noise from the most recent pressure sequence, thereby obtaining a filtered pressure sequence with a smooth change trend and improving the reliability and robustness of the detection results.

[0122] Specifically, the neural network model can use BP neural network model, convolutional neural network, RNN recurrent neural network, etc., wherein, taking BP neural network model as an example, model training can adopt nonlinear least square method, error algorithm can adopt mean square error method, front-end input layer can be a sequence of 10 consecutive air pressure difference values, the middle hidden layer is set to 15 layers, and the back-end output layer can be a classification result vector with the same number of floors as the building where the elevator is located. The training process of the neural network model can include: inputting the calibration air pressure data corresponding to the preset floor set into the constructed neural network model, wherein, except for the first floor as the reference air pressure value, the atmospheric pressure value of each floor needs to calculate the relative air pressure value with the reference air pressure value as the calibration air pressure data, calculate the error value according to the output prediction result and the set standard result, and update the model parameters based on the gradient descent method, and iterate until the error value reaches the preset threshold range, for example, less than 1.0×10 -6 , and obtain the trained neural network model.

[0123] Optionally, the specific implementation process of determining the elevator floor location based on the prediction results in step S104 may include: inputting the air pressure difference sequence into the trained neural network model to obtain a confidence set corresponding to all floors; and updating the floors corresponding to elements in the confidence set whose confidence reaches a set threshold as the elevator floor location. The confidence level represents the reliability of the prediction result for each floor; the higher the confidence level, the greater the likelihood that the floor is the elevator floor location.

[0124] In the specific implementation process, taking the total number of elevator floors as 36 as an example, the baseline air pressure value P1 of the first floor can be measured, and the recent air pressure sequence obtained in real time is P[n], n∈{1,2,…,N}. The calculation formula of the air pressure difference is as follows:

[0125] Δf[n]=(P[n]-P1),n∈{1,2,…,N}

[0126] The value of each element in Δf[n] represents the pressure difference between the air pressure at the current elevator location and the air pressure on the first floor. For example, 10 consecutive air pressure difference values ​​on Δf[n] can be combined into an air pressure difference value sequence:

[0127] ΔAp={Δf n-9 ,Δf n-8 ,…,Δf n-1 ,Δf n}

[0128] The pressure difference sequence ΔAp is transmitted to the trained neural network model, and 36 confidence sets R = {r1, r2, ..., r 36}, where the sum of the confidences of all elements in set R is 100%, the floors corresponding to the elements in set R whose confidences reach the set threshold can be updated as the elevator floor positions, where the set threshold can be set to 95%, 90%, etc. based on the accuracy requirements. For example, taking the threshold as 95%, at time t0, according to the calculation results of the neural network model, the confidence of the 5th floor is 95%, and the sum of the confidences of the other floors is 5%. Then, since the confidence of the 5th floor reaches 95%, the elevator floor position can be updated to the 5th floor. For example, at t n At this moment, according to the calculation results of the neural network model, the confidence of the 5th floor is 80%, and the sum of the confidences of the remaining floors is 20%. At this time, since the confidence of no floor reaches the set threshold, no floor update is performed. The elevator floor position is still the calculation result of the previous moment, and the movement direction of the elevator can be shown based on the vertical movement state of the elevator. The output information shows that the elevator is leaving from the 5th floor or is about to arrive at the 5th floor.

[0129] Therefore, by obtaining multiple consecutive air pressure values ​​as reference data and combining them with the trained neural network model to perform floor prediction, the accuracy and reliability of elevator floor position prediction can be improved.

[0130] In one embodiment, based on the above embodiment, after determining the vertical motion state of the elevator, the following process is further included:

[0131] Send the elevator floor position to the system backend so that the system backend can update the elevator floor position display.

[0132] It is worth noting that after the system background receives the elevator floor position, it can inform passengers of the floor the elevator is located on through a digital display at the elevator end, making it convenient for passengers to choose a suitable elevator according to the floor display and wait. The service robot can also obtain the real-time location information of the elevator based on the elevator floor position to further plan the route and synchronize the elevator floor position to the system background.

[0133] In the above method, a recent acceleration sequence of a first preset length and a recent air pressure sequence of a second preset length are obtained during elevator operation. The recent acceleration sequence is then transformed in the frequency domain to obtain a target spectrum, which is then converted into a power spectrum. When the power value corresponding to the power spectrum jumps from a first power range to a second power range, real-time acceleration value recording is started and stopped until the real-time acceleration value falls within a preset critical range, thereby obtaining a target acceleration sequence. The target acceleration sequence is then compared with a set state template sequence to determine the vertical motion state of the elevator. Finally, the difference between the elements in the recent air pressure sequence and the reference air pressure value is calculated to obtain an air pressure difference sequence. This air pressure difference sequence is input into a trained neural network model to obtain a prediction result, and the elevator floor position is determined based on the prediction result. This method reduces random data errors and effectively addresses abnormal data fluctuations by acquiring the recent acceleration sequence and the recent air pressure sequence in real time. The elevator operating state is determined by frequency domain power changes in the acceleration sequence and the state template sequence, and floor position is predicted using the neural network model based on the air pressure difference. This effectively improves the accuracy of elevator state detection and provides a reliable state data reference.

[0134] Figure 7 A flowchart of another elevator status detection method provided in an embodiment of the present application is shown in FIG. Figure 7 As shown, the elevator status detection method specifically includes the following steps:

[0135] Step S201: Acquire a recent acceleration sequence of a first preset length and a recent air pressure sequence of a second preset length during elevator operation. The acceleration data corresponding to the recent acceleration sequence is collected in real time by an accelerometer installed in the elevator car, and the air pressure data corresponding to the recent air pressure sequence is collected in real time by a barometer installed in the elevator car.

[0136] Step S202: Perform frequency domain transformation on the most recent acceleration sequence to obtain a target spectrum, and convert the target spectrum into a power spectrum.

[0137] Step S203: When the power value corresponding to the power spectrum jumps from the first power zone to the second power zone, start recording the real-time acceleration value until the real-time acceleration value is within a preset critical range and stop recording to obtain a target acceleration sequence. The target acceleration sequence is compared with the set state template sequence to determine the vertical motion state of the elevator. The first power zone corresponds to the power value range when the elevator is in a stationary state, and the second power zone corresponds to the power value range when the elevator is in a variable speed motion state.

[0138] Step S204: Calculate the difference between the elements in the most recent air pressure sequence and the reference air pressure value to obtain an air pressure difference sequence, input the air pressure difference sequence into the trained neural network model to obtain a prediction result, and determine the elevator floor position based on the prediction result.

[0139] Step S205: When the elevator floor position is different from the elevator floor position determined at the previous moment, the elevator floor position is added to the recorded floor change list.

[0140] Step S206: The floor change list is checked based on the set inspection rules. If the inspection result is abnormal, the floor change list is saved to the log, and the reference floor is added to the queue of floors to be executed in the elevator to complete the update of the reference air pressure value. It is worth noting that under normal circumstances, due to abnormal operating conditions such as power outages, restarts, or maintenance, errors in floor detection may occur occasionally. In order to enhance the reliability of detection, error correction mechanisms in steps S205 and S206 can be added to the floor detection process. It is understandable that when the elevator is in normal operation, the floor position change should be continuous. If the elevator is on the 4th floor at time t0, then the corresponding elevator floor position is 4, and it is added to the floor change list. Since the detection rate is higher than the elevator movement rate, after one unit of time, that is, at time t1, the elevator may not have moved, or may not have moved to the next floor, then the output result value is still 4, and the result value is the same as the result value at the previous moment, and the floor change list is not updated. Until n units of time have passed, the elevator moves to the 5th floor, and the output result value is 5. At this time, the result value is different from the result value at the previous moment, and the result value can be added to the floor change list. After each update of the floor change list, the floor change list can be checked based on the set inspection rules. For example, the inspection rule expression is as follows:

[0141]

[0142] According to the above inspection rules, when the difference between two adjacent items in the floor change list does not meet 1, it is judged as an abnormal situation. For example, if the value of one of the elements in the floor change list is 2, then the value of its adjacent element should be 1 or 3. If the inspection result is abnormal, the floor change list can be saved to the log, and the reference floor can be added to the elevator's queue of floors to be executed to complete the update of the reference air pressure value. Specifically, after the reference floor is added to the elevator's queue of floors to be executed, the elevator will continue to traverse the floors in the order of the queue of floors to be executed until it traverses to the reference floor. The elevator will reach the reference floor, such as the 1st floor, and complete the calibration of the reference air pressure value.

[0143] Therefore, by setting up an abnormal error correction mechanism, abnormal situations of the elevator can be recorded in time, and countermeasures such as recalibration of the reference air pressure value can be taken. The floor change list saved in the log can facilitate back-end personnel to confirm abnormalities and shut down the elevator for inspection in a timely manner, thereby ensuring the normal operation of the elevator and improving the accuracy of the detection results.

[0144] Figure 8 This is a structural block diagram of an elevator status detection device provided in an embodiment of the present application. The device is configured to execute the elevator status detection method provided in the above embodiment and has the corresponding functional modules and beneficial effects of the execution method. Figure 8 As shown, the device specifically includes:

[0145] An acquisition module 101 is configured to acquire a recent acceleration sequence of a first preset length and a recent air pressure sequence of a second preset length during elevator operation, wherein acceleration data corresponding to the recent acceleration sequence is acquired in real time by an accelerometer installed in the elevator car, and air pressure data corresponding to the recent air pressure sequence is acquired in real time by a barometer installed in the elevator car;

[0146] The frequency domain power conversion module 102 is configured to perform frequency domain transformation on the most recent acceleration sequence to obtain a target spectrum, and convert the target spectrum into a power spectrum;

[0147] The motion state determination module 103 is configured to, when the power value corresponding to the power spectrum jumps from the first power zone to the second power zone, start recording the real-time acceleration value, and stop recording when the real-time acceleration value is within a preset critical range, thereby obtaining a target acceleration sequence, and comparing the target acceleration sequence with a set state template sequence to determine the vertical motion state of the elevator, wherein the first power zone corresponds to a power value range when the elevator is in a stationary state, and the second power zone corresponds to a power value range when the elevator is in a variable speed motion state;

[0148] The floor position determination module 104 is configured to calculate the difference between the elements in the most recent air pressure sequence and the reference air pressure value to obtain an air pressure difference sequence, input the air pressure difference sequence into the trained neural network model to obtain a prediction result, and determine the elevator floor position based on the prediction result.

[0149] As can be seen from the above scheme, during elevator operation, a recent acceleration sequence of a first preset length and a recent air pressure sequence of a second preset length are obtained. The recent acceleration sequence is then transformed in the frequency domain to obtain a target spectrum, which is then converted into a power spectrum. When the power value corresponding to the power spectrum jumps from a first power range to a second power range, real-time acceleration value recording is started and stopped until the real-time acceleration value falls within a preset critical range. This target acceleration sequence is then compared with a set state template sequence to determine the vertical motion state of the elevator. Finally, the difference between the elements in the recent air pressure sequence and the reference air pressure value is calculated to obtain an air pressure difference sequence. This air pressure difference sequence is input into a trained neural network model to obtain a prediction result, and the elevator floor position is determined based on the prediction result. This method reduces random data errors and effectively addresses abnormal data fluctuations by acquiring the recent acceleration sequence and the recent air pressure sequence in real time. The elevator operating state is determined by frequency domain power changes in the acceleration sequence and the state template sequence, and floor position is predicted using the neural network model based on the air pressure difference. This effectively improves the accuracy of elevator state detection and provides a reliable state data reference.

[0150] In a possible embodiment, the state template sequence includes an ascending state sequence; the motion state determination module 103 is configured to:

[0151] Calculating a first sequence distance value between the target acceleration sequence and the ascending state sequence;

[0152] When the first sequence distance value is within a preset distance range, determining that the vertical motion state of the elevator is an ascending state;

[0153] When the first sequence distance value exceeds the preset distance range, it is determined that the vertical movement state of the elevator is a descending state.

[0154] In a possible embodiment, the state template sequence includes a descending state sequence; the motion state determination module 103 is configured to:

[0155] Calculating a second sequence distance value between the target acceleration sequence and the descending state sequence;

[0156] When the second sequence distance value is within the preset distance range, determining that the vertical motion state of the elevator is a descending state;

[0157] When the second sequence distance value exceeds the preset distance range, it is determined that the vertical motion state of the elevator is an ascending state.

[0158] In one possible embodiment, the training process of the neural network model includes:

[0159] The calibrated air pressure data corresponding to the preset floor set is input into the constructed neural network model, the error value is calculated based on the output prediction result and the set standard result, and the model parameters are updated based on the gradient descent method. It is iterated until the error value reaches the preset threshold range to obtain a trained neural network model.

[0160] In one possible embodiment, the floor position determination module 104 is configured to:

[0161] Inputting the air pressure difference sequence into the trained neural network model to obtain a confidence set corresponding to all floors;

[0162] The floor corresponding to the element in the confidence set whose confidence reaches a set threshold is used as the elevator floor position.

[0163] In a possible embodiment, a sequence filtering module is further included, configured to:

[0164] Filtering the most recent air pressure sequence to obtain a filtered air pressure sequence;

[0165] Accordingly, the motion state determination module 104 is configured as follows:

[0166] The elements in the filtered air pressure sequence are calculated to have differences with the reference air pressure value to obtain an air pressure difference sequence.

[0167] In a possible embodiment, the first transmitting module is further included, configured to:

[0168] The vertical motion state is sent to the system background so that the system background updates the lifting state display of the elevator.

[0169] In a possible embodiment, the second sending module is further included, configured to:

[0170] After determining the elevator floor position based on the prediction result, the method further includes:

[0171] The elevator floor position is sent to the system backend so that the system backend updates the floor position display of the elevator.

[0172] In a possible embodiment, a floor error correction module is further included, configured to:

[0173] In the case where the elevator floor position is different from the elevator floor position determined at a previous moment, adding the elevator floor position to a recorded floor change list;

[0174] The floor change list is checked based on the set inspection rules. If the inspection result is abnormal, the floor change list is saved to the log, and the reference floor is added to the queue of floors to be executed of the elevator to complete the update of the reference air pressure value.

[0175] In a possible embodiment, a reference air pressure module is further included, configured to:

[0176] When the elevator floor position is a reference floor, air pressure data of a preset time period is obtained, and a median value of the air pressure data is recorded as a reference air pressure value.

[0177] Figure 9 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application is shown in FIG. Figure 9 As shown, the device includes a processor 201, a memory 202, an input device 203 and an output device 204; the number of processors 201 in the device can be one or more. Figure 9 In the embodiment, a processor 201 is used as an example; the processor 201, the memory 202, the input device 203 and the output device 204 in the device can be connected by a bus or other means. Figure 9 The example of a bus connection is taken. The memory 202, as a computer-readable storage medium, can be configured to store software programs, computer executable programs, and modules, such as the program instructions / modules corresponding to the elevator status detection method in the embodiment of the present application. The processor 201 executes various functional applications and data processing of the device by running the software programs, instructions, and modules stored in the memory 202, thereby realizing the above-mentioned elevator status detection method. The input device 203 can be configured to receive input digital or character information, and generate key signal input related to user settings and function control of the device. The output device 204 may include a display device such as a display screen.

[0178] The electronic device provided above can be used to execute the elevator status detection method provided in any of the above embodiments, and has corresponding functions and beneficial effects.

[0179] An embodiment of the present application further provides a non-volatile storage medium containing computer-executable instructions. When the computer-executable instructions are executed by a computer processor, the computer-executable instructions are configured to perform an elevator state detection method described in the above embodiment, which includes:

[0180] Acquire a recent acceleration sequence of a first preset length and a recent air pressure sequence of a second preset length during elevator operation, wherein acceleration data corresponding to the recent acceleration sequence is collected in real time by an accelerometer installed in the elevator car, and air pressure data corresponding to the recent air pressure sequence is collected in real time by a barometer installed in the elevator car;

[0181] Perform frequency domain transformation on the latest acceleration sequence to obtain the target spectrum, and then convert the target spectrum into power spectrum;

[0182] When the power value corresponding to the power spectrum jumps from the first power zone to the second power zone, the real-time acceleration value is recorded, and the recording is stopped until the real-time acceleration value is within a preset critical range. The target acceleration sequence is obtained, and the target acceleration sequence is compared with the set state template sequence to determine the vertical motion state of the elevator. The first power zone corresponds to the power value range of the elevator in the static state, and the second power zone corresponds to the power value range of the elevator in the variable speed motion state.

[0183] The difference between the elements in the most recent air pressure sequence and the reference air pressure value is calculated to obtain an air pressure difference sequence. The air pressure difference sequence is input into the trained neural network model to obtain a prediction result, and the elevator floor position is determined based on the prediction result.

[0184] Storage medium - any of various types of memory devices or storage devices. The term "storage medium" is intended to include: installation media, such as CD-ROMs, floppy disks, or tape drives; computer system memory or random access memory, such as DRAM, DDR RAM, SRAM, EDO RAM, Rambus RAM, etc.; non-volatile memory, such as flash memory, magnetic media (such as hard disks or optical storage); registers or other similar types of memory elements, etc. Storage media may also include other types of memory or combinations thereof. In addition, the storage medium may be located in the first computer system in which the program is executed, or it may be located in a different second computer system that is connected to the first computer system via a network (such as the Internet). The second computer system can provide program instructions to the first computer for execution. The term "storage medium" may include two or more storage media residing in different locations (e.g., in different computer systems connected via a network). The storage medium may store program instructions (e.g., embodied as a computer program) that can be executed by one or more processors.

[0185] Of course, the storage medium containing computer-executable instructions provided in the embodiment of the present application is not limited to the elevator status detection method described above, and the computer-executable instructions can also execute related operations in the elevator status detection method provided in any embodiment of the present application.

[0186] It is worth noting that in the embodiment of the above-mentioned elevator status detection device, the various units and modules included are only divided according to functional logic, but are not limited to the above-mentioned division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not configured to limit the scope of protection of the embodiments of this application.

[0187] It should be noted that the numbering of each step in this solution is only used to describe the overall design framework of this solution, and does not represent the necessary order relationship between the steps. On the basis that the overall implementation process conforms to the overall design framework of this solution, it belongs to the protection scope of this solution, and the order of precedence in the form of text when describing is not an exclusive limitation on the specific implementation process of this solution. Those skilled in the art should understand that the embodiments of the present application can be provided as methods, systems, or computer program products. In a typical configuration, the computing device includes one or more processors (CPU), input / output interface, network interface and memory. The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash memory (flash RAM). The memory is an example of a computer-readable medium.

[0188] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0189] Note that the above are only preferred embodiments of the present invention and the technical principles employed. Those skilled in the art will appreciate that the present invention is not limited to the specific embodiments herein, and that various obvious changes, readjustments, and substitutions are possible for those skilled in the art without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments and may include many other equivalent embodiments without departing from the scope of the present invention. The scope of the present invention is determined by the scope of the appended claims.

Claims

1. An elevator status detection method, characterized in that: include: Acquire a recent acceleration sequence of a first preset length and a recent air pressure sequence of a second preset length during elevator operation, wherein acceleration data corresponding to the recent acceleration sequence is collected in real time by an accelerometer installed in the elevator car, and air pressure data corresponding to the recent air pressure sequence is collected in real time by a barometer installed in the elevator car; Performing frequency domain transformation on the most recent acceleration sequence to obtain a target spectrum, and converting the target spectrum into a power spectrum; When the power value corresponding to the power spectrum jumps from the first power zone to the second power zone, recording of the real-time acceleration value is started, and recording is stopped until the real-time acceleration value is within a preset critical range, to obtain a target acceleration sequence, and the target acceleration sequence is compared with a set state template sequence to determine the vertical motion state of the elevator, wherein the first power zone corresponds to a power value range when the elevator is in a stationary state, and the second power zone corresponds to a power value range when the elevator is in a variable speed motion state; Calculating the difference between the elements in the most recent air pressure sequence and the reference air pressure value to obtain an air pressure difference sequence, inputting the air pressure difference sequence into the trained neural network model to obtain a prediction result, and determining the elevator floor position based on the prediction result; The state template sequence includes an ascending state sequence and a descending state sequence, and comparing the target acceleration sequence with the set state template sequence to determine the vertical motion state of the elevator includes: Calculating a first sequence distance value between the target acceleration sequence and the ascending state sequence; When the first sequence distance value is within a preset distance range, determining that the vertical motion state of the elevator is an ascending state; When the first sequence distance value exceeds the preset distance range, determining that the vertical motion state of the elevator is a descending state; Calculating a second sequence distance value between the target acceleration sequence and the descending state sequence; When the second sequence distance value is within the preset distance range, determining that the vertical motion state of the elevator is a descending state; When the second sequence distance value exceeds the preset distance range, it is determined that the vertical motion state of the elevator is an ascending state.

2. The elevator state detection method according to claim 1, characterized in that: The training process of the neural network model includes: The calibrated air pressure data corresponding to the preset floor set is input into the constructed neural network model, the error value is calculated based on the output prediction result and the set standard result, and the model parameters are updated based on the gradient descent method. It is iterated until the error value reaches the preset threshold range to obtain a trained neural network model.

3. The elevator status detection method according to claim 1, characterized in that: The determining the elevator floor position based on the prediction result includes: Inputting the air pressure difference sequence into the trained neural network model to obtain a confidence set corresponding to all floors; The floor corresponding to the element in the confidence set whose confidence reaches the set threshold is updated as the elevator floor position.

4. The elevator state detection method according to claim 1, characterized in that: Before calculating the difference between the elements in the most recent air pressure sequence and the reference air pressure value to obtain the air pressure difference sequence, the method further includes: Filtering the most recent air pressure sequence to obtain a filtered air pressure sequence; Accordingly, the step of performing difference calculation between the elements in the most recent air pressure sequence and the reference air pressure value to obtain an air pressure difference sequence includes: The elements in the filtered air pressure sequence are calculated to have differences with the reference air pressure value to obtain an air pressure difference sequence.

5. The elevator status detection method according to claim 1, characterized in that: After determining the vertical motion state of the elevator, the method further includes: The vertical motion state is sent to the system background so that the system background updates the lifting state display of the elevator.

6. The elevator state detection method according to claim 1, characterized in that: After determining the elevator floor position based on the prediction result, the method further includes: The elevator floor position is sent to the system backend so that the system backend updates the floor position display of the elevator.

7. The elevator state detection method according to claim 1, characterized in that: The elevator state detection method further includes: In the case where the elevator floor position is different from the elevator floor position determined at a previous moment, adding the elevator floor position to a recorded floor change list; The floor change list is checked based on the set inspection rules. If the inspection result is abnormal, the floor change list is saved to the log, and the reference floor is added to the queue of floors to be executed of the elevator to complete the update of the reference air pressure value.

8. The elevator status detection method according to claim 1, characterized in that: The elevator state detection method further includes: When the elevator floor position is a reference floor, air pressure data of a preset time period is obtained, and a median value of the air pressure data is recorded as a reference air pressure value.

9. Elevator status detection device, characterized in that: include: an acquisition module configured to acquire a recent acceleration sequence of a first preset length and a recent air pressure sequence of a second preset length during elevator operation, wherein acceleration data corresponding to the recent acceleration sequence is acquired in real time by an accelerometer installed in the elevator car, and air pressure data corresponding to the recent air pressure sequence is acquired in real time by a barometer installed in the elevator car; a frequency domain power conversion module configured to perform frequency domain transformation on the most recent acceleration sequence to obtain a target spectrum, and convert the target spectrum into a power spectrum; a motion state determination module configured to, when the power value corresponding to the power spectrum jumps from a first power zone to a second power zone, start recording real-time acceleration values, stop recording until the real-time acceleration value is within a preset critical range, obtain a target acceleration sequence, compare the target acceleration sequence with a set state template sequence, and determine the vertical motion state of the elevator, wherein the first power zone corresponds to a power value range when the elevator is in a stationary state, and the second power zone corresponds to a power value range when the elevator is in a variable speed motion state; a floor position determination module configured to calculate the difference between the elements in the most recent air pressure sequence and the reference air pressure value to obtain an air pressure difference sequence, input the air pressure difference sequence into the trained neural network model to obtain a prediction result, and determine the elevator floor position based on the prediction result; The state template sequence includes an ascending state sequence and a descending state sequence, and the motion state determination module is configured as follows: Calculating a first sequence distance value between the target acceleration sequence and the ascending state sequence; When the first sequence distance value is within a preset distance range, determining that the vertical motion state of the elevator is an ascending state; When the first sequence distance value exceeds the preset distance range, determining that the vertical motion state of the elevator is a descending state; Calculating a second sequence distance value between the target acceleration sequence and the descending state sequence; When the second sequence distance value is within the preset distance range, determining that the vertical motion state of the elevator is a descending state; When the second sequence distance value exceeds the preset distance range, it is determined that the vertical motion state of the elevator is an ascending state.

10. An electronic device, comprising: one or more processors; The storage device is configured to store one or more programs, and when the one or more programs are executed by the one or more processors, the one or more processors implement the elevator state detection method according to any one of claims 1 to 8.

11. A non-volatile storage medium storing computer-executable instructions, wherein the computer-executable instructions are configured to execute the elevator state detection method according to any one of claims 1 to 8 when executed by a computer processor.

Images