Elevator position detection method based on impedance characteristics of traveling cable

Abstract

The elevator position detection method based on the impedance characteristics of the accompanying cable provided by the application realizes original absolute positioning through the mechanical-magnetic coupling characteristics of the cable itself, breaks away from the shackles of the traditional elevator positioning which relies on additional sensor devices, realizes the non-mechanical contact type measurement structure, greatly reduces the hardware failure rate and the later operation and maintenance cost of the elevator positioning system, and improves the long-term operation reliability of the system. In addition, by fusing the multi-dimensional position feature vectors such as the winding phase difference, the spectrum amplitude, the instantaneous frequency offset integral and the wireless channel, the feature discrimination of different height positions in the elevator shaft is greatly enriched, the problem of low recognition degree of similar position features in the shaft is effectively solved, the positioning error caused by the interference of electromagnetic noise and mechanical vibration on a single feature is inhibited, and the accuracy and robustness of the position feature representation are significantly improved.

Description

Technical Field

[0001] This invention relates to the field of elevator system position detection and positioning technology, and in particular to an elevator position detection method based on the impedance characteristics of the traveling cable. Background Technology

[0002] Real-time position detection of the elevator car within the shaft is a core prerequisite for ensuring safe operation, achieving precise leveling, and efficient scheduling. However, traditional solutions have long relied on magnetic shielding plates and light-blocking plates installed on the shaft walls in conjunction with car sensors to confirm position. This approach suffers from inherent drawbacks such as high installation workload, easy wear and tear on mechanical parts, and decreased reliability due to oil and dust contamination. While relative positioning methods such as rotary encoders can calculate displacement, they suffer from cumulative errors and are prone to losing position information after power failure, still requiring reset via absolute position correction switches within the shaft. In recent years, IoT technology has widely deployed wireless communication modules in modern elevators, but existing position detection solutions still rely on additional hardware within the shaft.

[0003] Among existing technologies, the more advanced solutions mainly fall into two categories. One category is position detection based on hoistway markers, where magnetic strips, RFID tags, or barcodes are installed at corresponding locations on each floor, and identified by readers on the elevator car. Compared to traditional mechanical switches, this solution offers improvements in accuracy and lifespan, but it remains essentially an add-on solution with a cumbersome installation process, and the markers are susceptible to failure due to vibration or oil contamination. The other category is absolute position measurement based on coded steel strips, which achieve millimeter-level positioning by laying coded steel strips along the entire length of the hoistway and using Hall effect sensors. However, the steel strip installation requires extremely high precision, is expensive, and mechanical deformation during long-distance operation can affect the reliability of readings. Overall, existing technologies still generally suffer from strong reliance on hoistway markers, inconsistent resolution across the entire speed range due to fixed sampling rates, and map table failure caused by cable aging and temperature drift during long-term operation.

[0004] Therefore, there is an urgent need to propose an elevator position detection method based on the impedance characteristics of the traveling cable. This method does not require the installation of additional markers or sensors in the shaft, nor does it alter the existing mechanical structure of the elevator. It utilizes only the elevator's original traveling cable and wireless communication module to achieve highly reliable, centimeter-level accuracy car position detection with consistent resolution across the entire speed range and long-term anti-drift stability, providing a practical and feasible technical path for the intelligent upgrading of elevators. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides an elevator position detection method based on the impedance characteristics of the traveling cable.

[0006] To achieve the above objectives, the technical solution of this invention is as follows: In a first aspect, the present invention provides an elevator position detection method based on the impedance characteristics of a traveling cable, wherein the traveling cable forms an absolute position anchor point related to the car height at a fixed mechanical constraint point due to impedance mismatch; the method includes: A high-frequency sweep signal is injected into the traveling cable at the elevator machine room end, and the full-band reflection signal of the high-frequency sweep signal is received in real time at the elevator car end via broadband radio frequency. Based on the processing of the full-band reflected signal, a multi-dimensional position feature vector is constructed, which includes the entanglement phase difference feature between the reflected wave and the incident wave, the reflection spectrum amplitude feature, the instantaneous frequency offset integral feature, and the wireless channel feature. The output rate of the multidimensional position feature vector is dynamically adjusted according to the real-time running speed of the car, and the multidimensional position feature vectors of the current time and multiple consecutive times before the current time are arranged in sequence to form a real-time feature sequence. An analytical mapping relationship between characteristic impedance and car height is established based on the mechanical morphology of the traveling cable, and an elevator shaft characteristic-height mapping table is established based on the analytical mapping relationship. The real-time feature sequence is matched with each candidate sequence in the elevator shaft feature-height mapping table to obtain a coarse positioning height. Interpolation refinement is performed in the neighborhood of the coarse positioning height, and post-processing refinement is performed in combination with the parsing mapping relationship as a prior constraint to output the car height value.

[0007] In some embodiments, injecting a high-frequency sweep signal into the traveling cable at the elevator machine room end includes: using an external signal coupling device at the machine room end, injecting a high-frequency sweep signal with a frequency range of 1MHz-1GHz into the traveling cable through broadband transformer isolation coupling or capacitive coupling; the high-frequency sweep signal avoids the communication frequency band of the elevator control system and the frequency band occupied by the operator's wireless equipment; at the elevator car end, a broadband radio frequency transceiver platform integrating a high-speed ADC and FPGA is used to directly sample the high-frequency sweep signal at a sampling rate of not less than 2GSPS, and the FPGA performs real-time FFT spectrum analysis to obtain the full-band reflected signal.

[0008] In some embodiments, the phase difference feature is represented as: In the formula, The characteristic of the entangled phase difference is that... Let be the signal propagation constant. The current equivalent transmission length of the accompanying cable. The inherent phase shift; the amplitude characteristics of the reflection spectrum are scattering parameters. frequency domain amplitude spectrum The amplitude value at the resonant frequency; the instantaneous frequency offset integral characteristic is expressed as: In the formula, This is the integral characteristic of instantaneous frequency offset. For instantaneous spectral shift, For integration variables, For the current moment, At the initial moment, The initial position equivalent transmission length of the traveling cable; the wireless channel feature is the RSSI (Radio Signal Strength Indicator) at the location of the elevator car; the multidimensional position feature vector is... .

[0009] In some embodiments, the step of dynamically adjusting the output rate of the multi-dimensional position feature vector based on the acquired real-time car speed includes: acquiring elevator acceleration and speed data in real time using a six-axis inertial measurement unit integrated at the car end; using a preset spatial displacement threshold as a trigger condition, outputting a feature vector when the IMU integral displacement reaches the threshold value, so that the spatial sampling interval of the elevator is constant during operation across the entire speed range; the instantaneous frequency offset integral feature is obtained by integrating the continuous phase over time after phase unwrapping, and an initial phase difference is assigned based on the initial position feature vector during power failure and restart to ensure the continuity of displacement integration.

[0010] In some embodiments, establishing the analytical mapping relationship between characteristic impedance and car height based on the mechanical morphology of the traveling cable includes: dividing the traveling cable along the hoistway height direction into a suspended section, a curved section, and a horizontal section determined by the car height; establishing a segmented leakage waveguide model of the traveling cable; and establishing the cable tension-impedance coupling equation. In the formula, Let be the characteristic impedance of the i-th segment of the traveling cable. This is the nominal characteristic impedance of the accompanying cable. and These are the nominal inductance and nominal capacitance per unit length of the accompanying cable, respectively. The tension of the i-th segment of the traveling cable The resulting change in inductance Car height The change in capacitance caused by bending deformation; the analytical mapping relationship between the characteristic impedance of each cable segment and the car height is established through the segmented leakage waveguide model and the cable tension-impedance coupling equation.

[0011] In some embodiments, the elevator shaft feature-height mapping table is pre-established through the following offline calibration steps: using an external high-precision measuring device as a true height reference source, the lowest floor level position of the elevator is defined as the physical height zero point; controlling the elevator to run from the bottom floor to the top floor at a preset uniform speed, while simultaneously detecting the car displacement using the external high-precision measuring device; whenever the car displacement reaches a preset spatial sampling interval, triggering the acquisition and recording of the multi-dimensional position feature vector corresponding to the current height; filtering and interpolating the multi-dimensional position feature vectors of each height sampling point to establish the elevator shaft feature-height mapping table in which the physical height corresponds one-to-one with the multi-dimensional position feature vector.

[0012] In some embodiments, the step of performing sequence similarity matching between the real-time feature sequence and each candidate sequence in the elevator shaft feature-height mapping table to obtain a coarse positioning height includes: using a preset neighborhood range of the previous positioning result as a search window in the feature-height mapping table, and truncating each candidate sequence within the search window by the sequence length of the real-time feature sequence; calculating the sequence distance between the real-time feature sequence and each of the candidate sequences using a dynamic time warping algorithm; and selecting the height interval corresponding to the candidate sequence with the smallest sequence distance as the coarse positioning height.

[0013] In some embodiments, the step of interpolating and refining within the neighborhood of the coarse positioning height, and then combining the analytical mapping relationship as a priori constraint for post-processing refinement to output the car height value, includes: interpolating and refining the coarse positioning height using nearest neighbor interpolation or linear interpolation within the neighborhood corresponding to the coarse positioning height to obtain a preliminary refined height; and using a priori physical model constructed based on the segmented leaky waveguide model and the cable tension-impedance coupling equation as a constraint, post-processing and refining the preliminary refined height using physical information Gaussian process regression to output the car height value.

[0014] The elevator position detection method based on the impedance characteristics of the traveling cable provided by this invention achieves native absolute positioning through the mechanical-electromagnetic coupling characteristics of the cable itself. This eliminates the reliance on additional sensors in traditional elevator positioning systems, enabling measurement without mechanical contact. This significantly reduces the hardware failure rate and subsequent maintenance costs of the elevator positioning system, and improves its long-term operational reliability. Furthermore, by fusing multi-dimensional position feature vectors such as winding phase difference, spectral amplitude, instantaneous frequency offset integral, and wireless channel, the method greatly enriches the feature distinguishability of different heights in the elevator shaft. This effectively solves the problems of feature confusion and low recognition between similar positions in the shaft, suppresses positioning errors caused by electromagnetic noise and mechanical vibration interference of single features, and significantly improves the accuracy and robustness of position feature representation. Meanwhile, this invention dynamically and adaptively adjusts the feature vector output rate according to the real-time running speed of the elevator car and constructs a continuous real-time feature sequence, realizing uniform temporal sampling based on spatial displacement. This effectively solves the problems of low-speed data redundancy, high-speed feature sparsity, and uneven sampling under variable speed conditions that exist in traditional fixed frame rate sampling. While ensuring the real-time positioning of the elevator during high-speed operation, it reduces the ineffective computing power overhead under low-speed stationary conditions, and balances the dynamic positioning response speed with the system's computing power and power consumption. It is suitable for all operating conditions of the elevator, including start-stop, variable speed, constant speed, and stationary conditions. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the elevator position detection method based on the impedance characteristics of the traveling cable provided in this embodiment of the invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limitations on this invention. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0017] In the following description, references to "some embodiments" refer to a subset of all possible embodiments; however, it is understood that "some embodiments" may be the same or different subsets of all possible embodiments and may be combined with each other without conflict. Unless otherwise defined, all technical and scientific terms used in the embodiments of the invention have the same meaning as commonly understood by one of ordinary skill in the art to which the embodiments of the invention pertain. The terminology used in the embodiments of the invention is for the purpose of describing the embodiments of the invention only and is not intended to limit the invention.

[0018] The following describes an exemplary application of the elevator position detection device based on the impedance characteristics of the traveling cable according to embodiments of the present invention. The elevator position detection device based on the impedance characteristics of the traveling cable provided in this embodiment can be implemented as a terminal or as a server. In one implementation, the elevator position detection device based on the impedance characteristics of the traveling cable provided in this embodiment can be implemented as various types of terminals such as laptops, tablets, desktop computers, and mobile devices. In another implementation, the elevator position detection device based on the impedance characteristics of the traveling cable provided in this embodiment can also be implemented as a server. The server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms. The terminal and the server can be directly or indirectly connected via wired or wireless communication, and this embodiment of the present invention does not impose any limitations. The following will describe an exemplary application of the elevator position detection device based on the impedance characteristics of the traveling cable when it is used as a server.

[0019] This invention provides an elevator position detection method based on the impedance characteristics of the traveling cable. See [link to relevant documentation]. Figure 1 , Figure 1 This is a flowchart illustrating an elevator position detection method based on the impedance characteristics of a traveling cable, provided by an embodiment of the present invention. Figure 1 The steps shown are explained.

[0020] S110, a high-frequency sweep signal is injected into the traveling cable at the elevator machine room end, and the full-band reflection signal of the high-frequency sweep signal is received in real time at the elevator car end via broadband radio frequency.

[0021] It should be noted that in this invention, the traveling cable not only undertakes the conventional functions of power supply and signal transmission, but also serves as a carrier for the transmission and reflection of high-frequency sweep signals. By utilizing its own mechanical shape and characteristic impedance variation law with the car height, position sensing distance can be realized.

[0022] Here, the high-frequency sweep signal refers to a periodic electromagnetic wave signal whose frequency changes continuously and linearly or nonlinearly with time within a preset high-frequency band, and serves as the detection excitation signal for this invention. Unlike fixed-frequency signals, the sweep signal can cover a continuous frequency band, enabling the acquisition of cable reflection characteristics across the entire frequency band, thus avoiding the shortcomings of single-frequency signals, which have limited characteristics and are easily interfered with.

[0023] Here, the elevator car end refers to the end point on the side wall of the elevator car where the signal receiving and acquisition module is fixedly installed. It is the end point for receiving and reflecting high-frequency signals and is a dynamic signal end that moves vertically up and down synchronously with the car. Together with the fixed elevator machine room end, it forms a signal acquisition architecture of fixed transmission + dynamic reception.

[0024] Here, the full-band reflected signal is a complete electromagnetic wave signal waveform obtained by analyzing the full-band reflected information. It is the concrete form of the original acquired data, covering all frequency components of the swept frequency signal. Unlike a single-frequency reflected signal, it can fully reflect the current physical state and transmission channel characteristics of the accompanying cable.

[0025] S120, based on the processing of the full-band reflected signal, construct a multi-dimensional position feature vector including the entanglement phase difference feature between the reflected wave and the incident wave, the reflection spectrum amplitude feature, the instantaneous frequency offset integral feature, and the wireless channel feature.

[0026] Here, the multidimensional location feature vector is a set of multidimensional features constructed in the form of vector data to characterize the position state of the elevator car. It is the positioning feature data carrier of this invention, which is composed of four types of core features. Each dimension of the vector corresponds to an independent feature parameter, realizing the quantitative representation of location information.

[0027] Here, the winding phase difference characteristic refers to the variation of the phase difference between the incident high-frequency sweep signal and the cable reflected echo signal with frequency band and time, caused by the signal transmission path length and cable deformation. As the traveling cable moves up and down with the car, it undergoes stretching and winding deformation, and the phase difference changes synchronously and regularly, which is the core phase characteristic characterizing the cable length and car height.

[0028] Here, the amplitude characteristics of the reflection spectrum refer to the combined characteristics of the amplitude, amplitude fluctuation, and amplitude attenuation of the reflected signal at each frequency point across the entire frequency band. The longer the cable is stretched and the greater the deformation, the more obvious the attenuation of high-frequency signal reflection. The amplitude characteristics of the spectrum can intuitively reflect the real-time shape of the cable.

[0029] Here, the instantaneous frequency offset integral characteristic refers to the cumulative result of the integral of the instantaneous frequency offset of the reflected signal relative to the incident signal within a preset time or frequency band. It can eliminate interference from instantaneous frequency fluctuations, improve the stability of the characteristic data, and accurately reflect the continuous deformation state changes of the cable.

[0030] Here, wireless channel characteristics refer to the comprehensive channel parameter characteristics of the high-frequency transmission channel formed by the accompanying cable, such as channel loss, channel delay, channel attenuation coefficient, and multipath propagation characteristics. The shaft environment and cable deformation directly change the channel characteristics, which are important dimensions of assisted positioning.

[0031] S130, dynamically adjust the output rate of the multi-dimensional position feature vector according to the acquired real-time running speed of the car, and arrange the multi-dimensional position feature vectors of the current time and multiple consecutive times before the current time in sequence to form a real-time feature sequence.

[0032] Here, the real-time operating speed of the elevator car refers to the instantaneous vertical ascent and descent speed of the elevator car at the current moment. This speed can be collected in real time by the encoder and speed sensor of the elevator main control system, and is divided into upward speed, downward speed, and stationary speed. The faster the car speed, the more frequently the position state changes, and the output rate of the corresponding feature vector needs to be adapted synchronously.

[0033] Here, the output rate of the multidimensional position feature vector refers to the number of times the system outputs the multidimensional position feature vector per unit time, measured in frames per second or times per second. To dynamically adapt to the car's operating state, the output rate is increased during high-speed operation and decreased during low-speed / stationary operation, balancing real-time positioning performance with system computing power and power consumption.

[0034] Here, the real-time feature sequence refers to a time-series data sequence formed by arranging multi-dimensional position feature vectors from the current moment and multiple consecutive historical moments in chronological order. Unlike feature vectors from a single moment, the time-series sequence can reflect the continuous changing trend of the car's position, thus avoiding positioning deviations caused by single-point feature errors.

[0035] S140, establish an analytical mapping relationship between characteristic impedance and car height based on the mechanical shape of the traveling cable, and establish an elevator shaft characteristic-height mapping table based on the analytical mapping relationship.

[0036] Here, the mechanical characteristics of the traveling cable refer to the set of physical parameters exhibited by the elevator traveling cable as it moves up and down with the elevator car, including cable tensile length, bending curvature, sag radius, and degree of deformation. The tensile and sag characteristics of the cable completely differ depending on the car height, forming the physical basis for impedance changes.

[0037] Here, characteristic impedance is the inherent impedance parameter of the traveling cable as a high-frequency electromagnetic wave transmission line, representing the cable's ability to impede electromagnetic waves during high-frequency signal transmission. Changes in the cable's tensile length and bending shape directly alter the characteristic impedance value, and this impedance change has a fixed physical relationship with the car height.

[0038] Here, the analytical mapping relationship is derived through electromagnetic field theory, mechanics of materials, and cable deformation formula. The quantitative mathematical function relationship between the characteristic impedance of the traveling cable and the height of the elevator car is a precise theoretical mapping formula. It has the characteristics of being calculable, derivable, and highly accurate, and serves as a priori theoretical constraint for subsequent precise positioning.

[0039] Here, the elevator shaft feature-height mapping table is a pre-calibrated and stored structured data table based on analytical mapping relationships. The table contains the standard feature sequence and standard impedance parameters corresponding to each standard height position in the elevator shaft. That is, each car height corresponds to a unique set of cable feature data, which is the benchmark database for real-time feature sequence matching and positioning.

[0040] S150, perform sequence similarity matching between the real-time feature sequence and each candidate sequence in the elevator shaft feature-height mapping table to obtain a coarse positioning height; perform interpolation refinement in the neighborhood of the coarse positioning height, and combine the parsing mapping relationship as a priori constraint for post-processing refinement to output the car height value.

[0041] Here, sequence similarity matching refers to the process of calculating the similarity between a real-time feature sequence and each candidate sequence in a mapping table using a time-series sequence alignment algorithm, and then selecting the height data corresponding to the candidate sequence with the highest similarity. This is the core algorithm logic of coarse localization. Commonly used algorithms include dynamic time warping and cosine similarity calculation.

[0042] Here, the initial car height value obtained by sequence similarity matching is a low-precision positioning result at the interval level. It can only determine the approximate height range of the car and has a small amount of calibration error. It needs to be refined to improve accuracy.

[0043] Here, interpolation refinement refers to the process of continuously fitting discrete standard height feature data within the coarse positioning height neighborhood using numerical algorithms such as linear interpolation, quadratic interpolation, and spline interpolation to obtain feature parameters corresponding to continuous height positions, thereby compensating for the discretization defects of the calibration table and improving positioning accuracy.

[0044] Here, post-processing refinement refers to the final optimization process of interpolation results, which, based on interpolation refinement and combined with the physical constraints of analytical mapping relationships, performs error correction, noise filtering, and data calibration to eliminate matching errors, interpolation errors, and environmental interference errors.

[0045] Here, the car height value refers to the final high-precision vertical height value of the elevator car relative to the bottom of the shaft / reference position after coarse matching, interpolation refinement, and post-processing constraints, which is the final positioning result of this invention.

[0046] The elevator position detection method based on the impedance characteristics of the traveling cable provided by this invention achieves native absolute positioning through the mechanical-electromagnetic coupling characteristics of the cable itself. This eliminates the reliance on additional sensors in traditional elevator positioning systems, enabling measurement without mechanical contact. This significantly reduces the hardware failure rate and subsequent maintenance costs of the elevator positioning system, and improves its long-term operational reliability. Furthermore, by fusing multi-dimensional position feature vectors such as winding phase difference, spectral amplitude, instantaneous frequency offset integral, and wireless channel, the method greatly enriches the feature distinguishability of different heights in the elevator shaft. This effectively solves the problems of feature confusion and low recognition between similar positions in the shaft, suppresses positioning errors caused by electromagnetic noise and mechanical vibration interference of single features, and significantly improves the accuracy and robustness of position feature representation. Meanwhile, this invention dynamically and adaptively adjusts the feature vector output rate according to the real-time running speed of the elevator car and constructs a continuous real-time feature sequence, realizing uniform temporal sampling based on spatial displacement. This effectively solves the problems of low-speed data redundancy, high-speed feature sparsity, and uneven sampling under variable speed conditions that exist in traditional fixed frame rate sampling. While ensuring the real-time positioning of the elevator during high-speed operation, it reduces the ineffective computing power overhead under low-speed stationary conditions, and balances the dynamic positioning response speed with the system's computing power and power consumption. It is suitable for all operating conditions of the elevator, including start-stop, variable speed, constant speed, and stationary conditions.

[0047] In some embodiments, injecting a high-frequency sweep signal into the traveling cable at the elevator machine room end includes: using an external signal coupling device at the machine room end, injecting a high-frequency sweep signal with a frequency range of 1MHz-1GHz into the traveling cable through broadband transformer isolation coupling or capacitive coupling; the high-frequency sweep signal avoids the communication frequency band of the elevator control system and the frequency band occupied by the operator's wireless equipment; at the elevator car end, a broadband radio frequency transceiver platform integrating a high-speed ADC and FPGA is used to directly sample the high-frequency sweep signal at a sampling rate of not less than 2GSPS, and the FPGA performs real-time FFT spectrum analysis to obtain the full-band reflected signal.

[0048] Here, the external signal coupling device at the machine room end specifically refers to a non-intrusive signal coupling device that is independently installed on the side of the elevator machine room equipment, without requiring any destructive modifications such as stripping, cutting, or rewiring of the existing elevator traveling cable. This device is independent of the elevator's original control, power supply, and communication systems. It can be fixed to the starting end of the traveling cable through physical bonding or terminal overlap, and is specifically used to superimpose and inject high-frequency sweep excitation signals onto the traveling cable. This structure differs from built-in embedded signal transmission modules, featuring no modification required, easy installation and disassembly, and strong versatility, making it suitable for the modification and installation of traveling cables in various elevator models.

[0049] Here, broadband transformer isolation coupling is an electromagnetic isolation coupling method suitable for broadband signal transmission. It relies on the electromagnetic induction principle of a broadband high-frequency transformer to achieve non-contact injection of high-frequency sweep signals into the traveling cable. Specifically, the primary side of the transformer is connected to a high-frequency sweep signal source, and the secondary side is coupled to the traveling cable. Signals are transmitted through electromagnetic coupling, while the electrical isolation characteristics of the transformer isolate the signal source from the elevator's power supply and control systems.

[0050] Here, capacitive coupling refers to a coupling method that utilizes the characteristic of capacitors to block DC and pass AC signals to achieve signal injection. A high-frequency sweep signal is superimposed onto the traveling cable through a high-frequency coupling capacitor. This method blocks low-frequency DC and power frequency AC, allowing only the sweep signal of the preset high-frequency band to be transmitted to the traveling cable. It features a simple structure, compact size, and excellent high-frequency response characteristics.

[0051] Here, the 1MHz-1GHz high-frequency sweep signal is the preferred effective operating frequency band for this embodiment, and it is a frequency range optimized for the electromagnetic environment of the elevator shaft. This frequency band combines the advantages of low-frequency signal transmission attenuation and strong penetration with high-frequency signal resolution and rich feature details. The lower frequency limit of 1MHz avoids the problems of low positioning accuracy and poor feature recognition of ultra-low frequency signals; the upper frequency limit of 1GHz avoids the defects of ultra-high frequency signals attenuating too quickly and having insufficient transmission distance in the shaft, and fully covers the effective sensing frequency band corresponding to the deformation and impedance changes of the elevator traveling cable.

[0052] In this embodiment, an external signal coupling device is used at the machine room end. Utilizing a non-intrusive injection method based on broadband transformer isolation coupling or capacitive coupling, it eliminates the need to disassemble, cut, or modify the elevator's original traveling cables and control system, thus preserving the elevator's original hardware structure and safety specifications. Compared to traditional positioning solutions requiring wiring modifications and built-in signal transmission modules, this embodiment is convenient to install, has low modification costs, and strong versatility. It can be adapted to both newly installed elevators and the upgrading of existing older elevators, demonstrating strong engineering feasibility. This embodiment avoids the elevator's own control system communication frequency band and the operator's public wireless frequency band by using the high-frequency sweep signal band, eliminating signal crosstalk at the spectrum source. On one hand, this prevents the detection sweep signal from interfering with the elevator's core functions such as control, communication, and intercom, eliminating safety hazards such as elevator malfunction, signal errors, and fault reports, ensuring the absolute safety and stability of the elevator's original control system. On the other hand, it avoids interference from external public wireless signals on the detection signal of this embodiment, significantly improving the purity of the collected signal and reducing positioning errors caused by environmental noise.

[0053] In some embodiments, the phase difference feature is represented as: In the formula, The characteristic of the entangled phase difference is that... Let be the signal propagation constant. The current equivalent transmission length of the accompanying cable. The inherent phase shift; the amplitude characteristics of the reflection spectrum are scattering parameters. frequency domain amplitude spectrum The amplitude value at the resonant frequency; the instantaneous frequency offset integral characteristic is expressed as: In the formula, This is the integral characteristic of instantaneous frequency offset. For instantaneous spectral shift, For integration variables, For the current moment, At the initial moment, The initial position equivalent transmission length of the traveling cable; the wireless channel feature is the RSSI (Radio Signal Strength Indicator) at the location of the elevator car; the multidimensional position feature vector is... .

[0054] In some embodiments, the step of dynamically adjusting the output rate of the multi-dimensional position feature vector based on the acquired real-time car speed includes: acquiring elevator acceleration and speed data in real time using a six-axis inertial measurement unit integrated at the car end; using a preset spatial displacement threshold as a trigger condition, outputting a feature vector when the IMU integral displacement reaches the threshold value, so that the spatial sampling interval of the elevator is constant during operation across the entire speed range; the instantaneous frequency offset integral feature is obtained by integrating the continuous phase over time after phase unwrapping, and an initial phase difference is assigned based on the initial position feature vector during power failure and restart to ensure the continuity of displacement integration.

[0055] Here, the six-axis inertial measurement unit is a miniature inertial sensing module integrating a three-axis accelerometer and a three-axis gyroscope, integrated into the elevator car end, and belongs to the category of high-precision motion attitude acquisition devices. The three-axis accelerometer is used to collect linear acceleration data in the three-dimensional space of the car, and the three-axis gyroscope is used to collect angular velocity data along the three axes of the car. Through data fusion and integration, the instantaneous velocity, displacement, and motion state of the car can be calculated in real time. This invention mainly utilizes its vertical acceleration and velocity output capabilities to achieve high-frequency, independent acquisition of elevator car motion parameters, without relying on the elevator main control encoder signal.

[0056] Here, elevator acceleration and velocity data are dynamic parameters of the elevator car's vertical motion within the shaft. Elevator acceleration is the change in the car's vertical speed per unit time, including acceleration values ​​corresponding to four states: acceleration, deceleration, constant speed, and stationary. Elevator velocity data is the instantaneous rate of the car's real-time vertical motion. This set of data is the core control basis for dynamically adjusting the characteristic vector output rate in this embodiment.

[0057] Here, the preset spatial displacement threshold is a fixed spatial displacement threshold that is pre-calibrated and stored to trigger the output of multi-dimensional position feature vectors. It is a preset constant physical distance value, which is different from the time trigger threshold. For example, the spatial displacement threshold can be set to a fixed small distance such as 2mm or 5mm. That is, the system outputs a feature vector frame for each actual movement of the car corresponding to a fixed physical distance, regardless of the speed of the car.

[0058] Here, IMU integral displacement refers to the continuous motion displacement of the elevator car obtained by performing time-domain integration on the real-time acceleration and velocity data collected by the six-axis inertial measurement unit. IMU integral displacement can output the real-time movement distance of the car at high frequency and continuously, which serves as the quantitative basis for determining whether the feature output is triggered in this invention, realizing displacement measurement in pure spatial dimension, unaffected by electrical sampling clock jitter or running speed fluctuations.

[0059] Here, "constant spatial sampling interval" means that the physical distance of the elevator car corresponding to the multi-dimensional position feature vectors of two adjacent frames remains constant at any operating speed. Traditional timed sampling has a fixed time interval and the spatial interval varies with speed. This invention has a fixed spatial interval and the output frame rate adapts to speed, achieving uniform spatial sampling across the entire speed range.

[0060] Here, the phase unwrapping process is performed on the folded phase of the high-frequency reflected signal to obtain a phase value that is continuously changing without jumps. Conventional radio frequency reflected phases have periodic folding jumps in the range of [-π,π] or [0,2π], which cannot be directly used for integration calculations; phase unwrapping compensates for and reconnects the abrupt phase to obtain a true phase curve that changes monotonically and continuously with the car displacement. This curve is the continuous phase after unwrapping.

[0061] Here, the initial position feature vector refers to the multi-dimensional position feature vector that the system collects and constructs for the first time after each power-on restart of the elevator, which corresponds to the current physical position of the car. It includes all dimensional parameters such as phase features and spectral amplitude features corresponding to the current position, and can accurately represent the true position state of the car at the moment of restart.

[0062] Here, the initial phase difference refers to the initial phase offset reference value obtained by the system based on the initial position feature vector at the restart time and by assigning it. It is used to make up for the historical phase accumulation lost during the power outage, so that the starting point of the phase integration after restart is continuously aligned with the phase change trend before the power outage.

[0063] Here, the continuity of displacement integral refers to the fact that the displacement correlation parameters obtained by the system through time integration of instantaneous frequency offset and phase change maintain numerical continuity, without jumps or reset faults, under all working conditions of elevator normal operation, power failure restart, and start-stop speed change, and will not exhibit phenomena such as the integral baseline returning to zero, position jumps, or characteristic breaks due to power failure restart.

[0064] This embodiment independently acquires acceleration and velocity data through a six-axis inertial measurement unit at the car end, without relying on the elevator main control encoder, bus data, or signals from the elevator's native control system. This avoids characteristic rate adjustment failures caused by elevator main control signal delays, packet loss, or malfunctions. Furthermore, it enables autonomous sensing and control of the positioning system, improving its independence, compatibility, and anti-interference capabilities, and adapting to elevators of various brands and control protocols. Simultaneously, this embodiment uses unwound continuous phases for time integration, eliminating the abrupt integral changes and distortions caused by the periodic jumps of traditional folded phases. This allows the instantaneous frequency offset integral characteristics to accurately and linearly reflect the deformation of the traveling cable and the car's displacement, improving the stability and accuracy of the multi-dimensional position feature vector. This provides high-linearity, low-noise core feature parameters for subsequent sequence matching and height calculation.

[0065] In some embodiments, establishing the analytical mapping relationship between characteristic impedance and car height based on the mechanical morphology of the traveling cable includes: dividing the traveling cable along the hoistway height direction into a suspended section, a curved section, and a horizontal section determined by the car height; establishing a segmented leakage waveguide model of the traveling cable; and establishing the cable tension-impedance coupling equation. In the formula, Let be the characteristic impedance of the i-th segment of the traveling cable. This is the nominal characteristic impedance of the accompanying cable. and These are the nominal inductance and nominal capacitance per unit length of the accompanying cable, respectively. The tension of the i-th segment of the traveling cable The resulting change in inductance Car height The change in capacitance caused by bending deformation; the analytical mapping relationship between the characteristic impedance of each cable segment and the car height is established through the segmented leakage waveguide model and the cable tension-impedance coupling equation.

[0066] In some embodiments, the elevator shaft feature-height mapping table is pre-established through the following offline calibration steps: using an external high-precision measuring device as a true height reference source, the lowest floor level position of the elevator is defined as the physical height zero point; controlling the elevator to run from the bottom floor to the top floor at a preset uniform speed, while simultaneously using the external high-precision measuring device to detect the car displacement, and triggering the acquisition and recording of the multi-dimensional position feature vector corresponding to the current height whenever the car displacement reaches a preset spatial sampling interval; filtering and interpolating the multi-dimensional position feature vectors of each height sampling point to establish the elevator shaft feature-height mapping table in which the physical height corresponds one-to-one with the multi-dimensional position feature vector.

[0067] Here, the height true reference source refers to a benchmark data source that can output the absolute, accurate, and traceable true value of the elevator car's physical height. In this embodiment, the measurement data from an external high-precision measuring device is used as the height true reference source to benchmark and calibrate the feature data collected by this system, ensuring that the correspondence between the feature vector and the physical height is true and accurate. This is the core benchmark basis for the high-precision construction of the mapping table.

[0068] Here, the physical height zero point is a human-defined absolute height reference zero point of the elevator shaft. In this embodiment, it specifically refers to the lowest floor leveling position of the elevator, which is the standard physical position when the elevator car stops at the lowest floor leveling position. This zero point is the unified reference for measuring the height of the entire shaft. All car height values ​​are calculated by accumulating them based on this point, thereby achieving standardization and unified quantification of shaft height.

[0069] Here, the preset constant speed is a stable and unchanging elevator lifting speed set manually during the offline calibration phase. This speed is preferably low and stable to avoid instantaneous deformation and vibration interference of the cable caused by acceleration, deceleration, and start-stop jitter, ensuring that the changes in the shape and impedance of the accompanying cable are stable and regular during the calibration process, and improving the calibration characteristic numbers.

[0070] Here, the preset spatial sampling interval refers to a fixed physical distance interval pre-set during the offline calibration phase. It serves as the trigger condition for collecting calibration data and is a constant spatial scale parameter, independent of time and speed. For example, a fixed spatial interval of 1cm or 2cm can be set, meaning that for every fixed physical distance the car rises, the system synchronously collects a set of corresponding multi-dimensional position feature vectors, ensuring uniform distribution of sampling points throughout the hoistway.

[0071] In this embodiment, the system abandons the relative reference method of self-calibration and uses an external high-precision measuring device as the true reference source for height. Simultaneously, the bottom level is defined as a unified physical zero point, ensuring absolute physical traceability of the height data in the entire mapping table. This eliminates the reference offset and zero-point drift problems caused by self-calibration, fundamentally guaranteeing the reference accuracy of subsequent online positioning and ensuring no systematic deviation in long-term positioning. Furthermore, a uniform speed calibration operation mode is adopted to avoid interference from dynamic cable deformation caused by speed changes and start-stop conditions, ensuring stable cable status and reliable feature data during calibration. At the same time, fixed spatial sampling intervals are used for triggering acquisition, rather than fixed-time sampling, resulting in uniform sampling point density across the entire wellbore's elevation and gradation areas. This avoids problems such as sparse local features and redundant local data, ensuring the balance and integrity of positioning features throughout the entire travel range.

[0072] Meanwhile, this embodiment performs filtering on the original feature vector of calibration, which can effectively filter out random noise such as electromagnetic interference, mechanical vibration, and environmental clutter during the calibration process, eliminate abnormal and distorted data, smooth the fluctuation of the original feature, retain the real position features corresponding to the deformation of the accompanying cable, avoid noise data being entered into the mapping table, effectively improve the anti-interference ability and data reliability of the benchmark database, and reduce random errors in subsequent online matching and positioning.

[0073] In some embodiments, the step of performing sequence similarity matching between the real-time feature sequence and each candidate sequence in the elevator shaft feature-height mapping table to obtain a coarse positioning height includes: using a preset neighborhood range of the previous positioning result as a search window in the feature-height mapping table, and truncating each candidate sequence within the search window by the sequence length of the real-time feature sequence; calculating the sequence distance between the real-time feature sequence and each of the candidate sequences using a dynamic time warping algorithm; and selecting the height interval corresponding to the candidate sequence with the smallest sequence distance as the coarse positioning height.

[0074] Here, the previous positioning result refers to the precise car height value that the system successfully calculated and output in the previous calculation cycle, after refinement and correction. During the continuous operation of the elevator, the system continuously outputs height results at fixed or adaptive cycles. The car displacement is minimal at adjacent moments, and the previous positioning result can accurately represent the car's true position at the previous moment, providing effective prior position constraints for the current positioning search.

[0075] Here, the preset neighborhood range is a fixed or adaptive small height interval pre-configured based on the previous positioning height result, used to limit the effective range of feature retrieval at the current moment. For example, the preset neighborhood range can be set to ±5cm or ±10cm, that is, with the output height of the previous frame as the center, the range is formed by floating up and down by a fixed distance; this range is smaller than the maximum feasible displacement of the elevator in a single cycle, which can ensure that the neighborhood range completely includes the current real position of the car.

[0076] Here, the search window is a local data interval defined by the preset neighborhood range of the previous positioning result, used for candidate sequence retrieval in the elevator shaft feature-height mapping table. Unlike the global traversal search of the entire shaft mapping table, this embodiment only retrieves data within a narrow neighborhood interval, thus narrowing the matching range and reducing computational load.

[0077] Here, the candidate sequence with the smallest sequence distance refers to the standard feature sequence that has the highest consistency with the current real-time feature sequence in terms of feature trend, dimensional change, and time series deviation among all candidate sequences captured in the search window. The hoistway height position corresponding to this candidate sequence is closest to the current actual physical position of the car.

[0078] In this embodiment, leveraging the continuous operation and non-abrupt positional motion characteristics of the elevator, a local search window is constructed based on the accurate positioning result of the previous frame. This abandons the traditional method of global traversal matching across the entire shaft mapping table, significantly narrowing the search range from the entire shaft to a tiny neighborhood interval. This effectively reduces the number of candidate sequences to be matched, significantly lowers computational power consumption, shortens algorithm computation time, and ensures real-time positioning output under high-speed operation conditions, adapting to the dynamic high-speed positioning requirements of elevators. By truncating and regularizing the candidate sequences within the window according to the real-time feature sequence length, the dimensions, frame numbers, and temporal structure of the two sets of time-series data to be matched are completely aligned. This eliminates problems such as algorithm inability to compute, matching deviation, and feature misalignment caused by inconsistent sequence lengths, ensuring that the input conditions for each sequence similarity match are standardized and normalized, improving the stability and reliability of the matching process.

[0079] In addition, using quantified sequence distance as the matching criterion, the coarse positioning height is determined by selecting the candidate sequence with the smallest distance. Compared with traditional single feature comparison and threshold discrimination methods, the evaluation criteria are more objective and accurate. It can effectively screen out the standard position interval that best fits the real-time running state, effectively avoid the problem of mismatch and skip match caused by noise interference, and provide high-precision and high-reliability coarse positioning initial values ​​for subsequent refined positioning.

[0080] In some embodiments, the step of interpolating and refining within the neighborhood of the coarse positioning height, and then combining the analytical mapping relationship as a priori constraint for post-processing refinement to output the car height value, includes: interpolating and refining the coarse positioning height using nearest neighbor interpolation or linear interpolation within the neighborhood corresponding to the coarse positioning height to obtain a preliminary refined height; and using a priori physical model constructed based on the segmented leaky waveguide model and the cable tension-impedance coupling equation as a constraint, post-processing and refining the preliminary refined height using physical information Gaussian process regression to output the car height value.

[0081] Here, the preliminary refined height refers to the optimized height result obtained after neighborhood interpolation refinement. Compared with the coarse positioning height, this height has broken free from the limitations of discrete calibration points, possesses continuous height resolution, and eliminates the inherent errors caused by discrete sampling step sizes. However, it has not yet incorporated physical model constraints, and still contains a small amount of algorithm fitting error and environmental noise error, requiring further post-processing correction.

[0082] Here, the segmented leakage waveguide model is a physical simulation model established for the high-frequency signal transmission characteristics of the elevator traveling cable, and it is one of the core prior physical models in this embodiment. As the elevator car rises and falls, the traveling cable exhibits segmented suspension, bending, and tension, resulting in leakage radiation and transmission attenuation of high-frequency radio frequency signals along the cable. This model divides the entire traveling cable into multiple segments according to its suspension shape and tension distribution, and establishes the physical calculation relationships for electromagnetic wave leakage, transmission loss, and phase delay in each segment, accurately describing the high-frequency transmission characteristics of the cable at different car heights.

[0083] Here, the cable tension-impedance coupling equation is a mathematical equation characterizing the quantitative coupling relationship between the mechanical tension of the traveling cable and its high-frequency characteristic impedance, and it serves as the core analytical constraint equation in this embodiment. Changes in elevator car height alter the cable sag length and tension, leading to cable dielectric deformation, changes in line spacing, and ultimately, a regular shift in characteristic impedance. This equation, derived through mechanical-electromagnetic coupling, obtains a rigorous mathematical mapping relationship between tension, deformation, impedance, and height, providing a physical truth constraint for the positioning results.

[0084] This embodiment is compatible with both nearest neighbor and linear interpolation, adaptively switching based on the characteristic linearity of each section of the shaft: nearest neighbor interpolation is used in stable sections to reduce computational load and suppress numerical jitter; linear interpolation is used in sections with significant changes to improve continuous fitting accuracy, effectively eliminating the step-like positioning error caused by discrete calibration points, smoothing coarse positioning into continuous height values, and achieving an initial leap in accuracy. Based on this, a segmented leaky waveguide model is introduced to characterize high-frequency transmission loss, and a complete physical prior constraint system is constructed by relating mechanical deformation and electromagnetic parameter changes through the cable tension-impedance coupling equation. This limits erroneous solutions and pseudo-extreme solutions from a mechanistic perspective, compensating for the shortcomings of pure data-driven algorithms, such as susceptibility to environmental interference and lack of boundary constraints. The physical information Gaussian process regression integrates measured data and physical laws, adaptively fitting the true position offset while forcing the results to conform to the electromagnetic and mechanical changes of the cable. This effectively filters out shaft noise and vibration random errors, corrects residual system biases from interpolation, and further improves the smoothness and accuracy of the output. The system forms a step-by-step error convergence path from coarse matching to continuous interpolation to physical constraint refinement, which reduces the risk of single-fitting and can maintain high-precision and stable output even under complex working conditions such as elevator speed change, vibration, electromagnetic interference, and temperature deformation.

[0085] The following will describe an exemplary application of the embodiments of the present invention in a practical application scenario.

[0086] To address the reliance on additional markers or sensors within the hoistway in existing elevator position detection technologies, a novel method is designed that utilizes the elevator's inherent traveling cable as a distributed sensor. This method eliminates the need for additional markers or sensors within the hoistway and does not alter the elevator's mechanical structure. It combines the unique electromagnetic impedance characteristics generated by the cable at different floor positions due to fixed mechanical constraints with radio frequency environmental signatures (RSSI) obtained from the SDR receiver module at the car end to achieve highly reliable, centimeter-level accuracy elevator car position detection. The method includes the following: 1) Hardware deployment and physical layer construction: This embodiment utilizes the elevator's existing traveling cable as a distributed sensor. A signal injection module is deployed in the machine room to inject a 1MHz-1GHz high-frequency sweep signal into the cable. This module is not connected to the safety circuit and integrates a spectrum sensing unit to monitor the wireless frequency band occupancy in the hoistway in real time, dynamically avoiding co-channel interference. The module, along with the car-side SDR receiver module and 4G / 5G DTU communication module, all possess the National Radio Transmission Equipment Type Approval Code (CMIIT ID). Powered by the elevator's main power supply, it is also connected to the emergency lighting power supply as a backup (utilizing the existing emergency circuit, eliminating the need for a new UPS), ensuring that detection can continue for a preset duration even after a mains power outage.

[0087] The car end adopts a broadband direct sampling architecture, equipped with a small RF receiver module (integrating a high-speed ADC and FPGA). It picks up full-band reflected signals through capacitive coupling and performs direct sampling at a sampling rate of no less than 2 GSPS. The FPGA performs FFT spectrum analysis in real time. It also integrates a six-axis IMU (SPI / I2C interface, ≥50Hz real-time velocity / acceleration data). The 4G / 5G DTU is only used for wireless backhaul of feature data and positioning results, realizing communication link multiplexing rather than sharing the RF front-end.

[0088] Based on this, a segmented waveguide model of the traveling cable was established. Combining the tension-impedance coupling equation, the mechanical morphology of the cable (suspended section, bent section, horizontal section) was correlated with the characteristic impedance distribution, clarifying the impedance-height coupling mechanism. Considering the batch differences in cable materials and the discreteness of installation processes, which make it difficult to accurately calibrate the model parameters online, a data-driven feature-height mapping table was used for actual positioning, constructing a physical layer foundation that can sense the car height.

[0089] 2) Signal acquisition and feature extraction: The car-end broadband RF receiver synchronously monitors the high-frequency sweep signal and uses the broadband direct sampling architecture described above to acquire digital spectrum information. After digital down-conversion and anti-aliasing filtering, the FPGA performs real-time FFT spectrum analysis. Based on this, the sampled signal undergoes 50Hz power frequency filtering, digital down-conversion, and anti-aliasing filtering to extract the entangled phase difference between the reflected and incident waves. , reflection spectrum amplitude Instantaneous frequency offset integral Furthermore, it integrates the radio channel features (RSSI) obtained by scanning with the broadband radio frequency receiver module to construct a four-dimensional location feature vector. The FPGA continuously outputs FFT spectrum analysis results at a fixed high frame rate; the feature extraction module adaptively decides whether to output a four-dimensional feature vector based on whether the IMU integral displacement reaches the 1cm threshold, so that the spatial sampling interval is constant at 1cm, thus completing the conversion from electromagnetic signal to quantifiable feature.

[0090] 3) Establish an elevator shaft feature-height mapping table and an adaptive maintenance mechanism: A laser rangefinder was used as a temporary offline calibration tool, with the laser rangefinder / rotary encoder serving as the true reference. The ground floor level was defined as the zero height point. The elevator moved upwards at a constant speed of 0.2-0.5 m / s, triggering feature vector acquisition every 1 cm of displacement. The same height point was repeatedly sampled, and after median filtering to remove outliers, the average was taken to generate the standard feature vector corresponding to that height. Subsequently, cubic spline interpolation was used to complete invalid data, establishing a physical height-four-dimensional feature vector mapping database.

[0091] To enhance environmental adaptability, the EWMA algorithm is introduced to achieve dynamic and smooth tracking of the elevator shaft feature-height mapping table. Simultaneously, existing temperature and humidity sensors within the elevator car are used to establish a humidity-impedance drift compensation model to correct impedance shifts caused by cable dampness due to conditions such as pit water accumulation. Combined with a temperature compensation mechanism, this enables adaptive correction of the elevator shaft feature-height mapping table to combined temperature and humidity changes. Furthermore, an adaptive recalibration mechanism for cable replacement is designed: for replacements of cables of the same specification, a fast correction strategy updates the mapping parameters; for replacements of cables of different specifications, a full-stroke recalibration is triggered to ensure the long-term stability and reliability of the mapping database. This training strategy effectively balances accuracy, robustness, and engineering maintainability.

[0092] 4) Online operation phase: Real-time acquisition of the current four-dimensional feature vector (i.e., the multi-dimensional positional feature vector in the above embodiment) X, and extraction of the most recent N consecutive moments (N=10-50, adaptively adjusted with speed) to form a feature sequence. Then, candidate sequences are truncated to the elevator shaft feature-height mapping table with the same window length. The algorithm employs Sakoe-Chiba Band Dynamic Time Warping (DTW) to limit the search scope to calculating sequence similarity within the coarse localization neighborhood, selecting the candidate sequence with the smallest distance corresponding to the height interval. Finally, in the online stage, lookup table interpolation (such as nearest neighbor interpolation) is used to refine the localization results, and with the physical model as a prior constraint, Gaussian process regression based on physical information is used for post-processing refinement to improve generalization ability, outputting centimeter-level car height, which can be further mapped to the corresponding floor.

[0093] In some implementations, the hardware deployment and physical layer construction described above in 1) can also be achieved in the following ways: This embodiment establishes a segmented waveguide model of the traveling cable in the shaft, dividing the cable along the shaft height into a suspension segment, a bending segment, and a horizontal segment. Each segment exhibits different characteristic impedance distributions due to differences in gravity suspension configuration and the bending radius of fixed mechanical constraint points such as cable racks / hangers. Let the characteristic impedance of the i-th segment of the traveling cable be... , The height of the car. Let be the tension of the i-th segment of the accompanying cable. Simultaneously, establish the cable tension-impedance coupling equation. , Let be the characteristic impedance of the i-th segment of the traveling cable. This is the nominal characteristic impedance of the accompanying cable. and These are the nominal inductance and nominal capacitance per unit length of the accompanying cable, respectively. The tension of the i-th segment of the traveling cable The resulting change in inductance Car height The change in capacitance caused by bending deformation.

[0094] This formula establishes an analytical relationship between the characteristic impedance of each segment and the car height, used to explain the impedance-height coupling mechanism. However, due to batch differences in cable materials and the discreteness of installation processes, the model parameters are difficult to calibrate accurately online. Therefore, the actual positioning uses a data-driven elevator shaft characteristic-height mapping table instead of directly performing model inverse kinematics. At fixed mechanical constraint points (such as cable racks and brackets), the cable experiences inherent impedance mismatch due to the fixed bending radius, forming periodic resonant frequency points or deep fading points. These frequency points are strongly correlated with the car height and can serve as absolute position anchor points, forming the basis of positioning in this embodiment. As the car moves up and down in the shaft, the effective transmission length of the cable continuously changes, and the phase and amplitude of the reflected signal exhibit a regular response accordingly. In addition, the signal injection module, the car-end SDR receiving module, and the 4G / 5G DTU module are all powered by the elevator main power supply and connected to the emergency lighting power supply as a backup (using the existing emergency circuit, no additional UPS configuration is required), ensuring that the positioning function can still maintain the preset duration after the mains power is interrupted.

[0095] In some implementations, the signal acquisition and feature extraction described in 2) above can also be achieved in the following ways: The broadband RF receiving module at the car end synchronously monitors the high-frequency sweep signal injected from the machine room, while the 4G / 5G DTU module maintains conventional wireless communication functions. The RF layer directly reuses the aforementioned broadband direct sampling architecture, with a high-speed ADC completing full-band digital acquisition at a sampling rate of no less than 2GSPS. After real-time FFT spectrum analysis by the FPGA, the reflection amplitude and phase information of each frequency point are extracted to synthesize the full-band reflection spectrum. The FPGA continuously outputs the FFT spectrum analysis results at a fixed high frame rate (e.g., 50Hz). The feature extraction module adaptively decides whether to output a four-dimensional feature vector based on whether the IMU integral displacement reaches a 1cm threshold. Specifically, the IMU integrates the car displacement in real time, triggering a feature vector output every time the cumulative displacement reaches 1cm, keeping the spatial sampling interval constant at 1cm. If the elevator speed varies within the range of 0.2m / s to 0.5m / s, the feature output rate adaptively varies between 20Hz and 50Hz accordingly. The fixed frame rate limit of the FPGA is constrained by the single FFT processing cycle (i.e. ≤50Hz). If the speed exceeds the limit, the feature extraction module will still output according to the IMU 1 cm threshold. The undersampled part will be recovered to the 1 cm equivalent resolution by DTW sequence matching and table lookup interpolation compensation, ensuring that the spatial resolution is constant throughout the entire speed range.

[0096] When a high-frequency signal is transmitted in the traveling cable, changes in the car's position alter the cable's impedance characteristics, causing a regular shift in the phase difference between the reflected and incident waves and in the amplitude of the reflected spectrum. Let the current equivalent transmission length of the traveling cable be S, and the signal angular frequency be... The propagation constant is , Let be the signal propagation constant, then the entanglement phase of the reflected signal at the characteristic frequency is: , Due to the inherent phase shift, the winding phase maps the total phase accumulation to the interval [0, 2π); since the cable has multiple periodic mechanical constraint anchor points, the total reflected signal is a superposition of multiple reflection vectors, using scattering parameters Describe the multiple reflection effect and extract its frequency domain amplitude spectrum. The amplitude value at the resonant frequency is taken as the amplitude characteristic of the reflection spectrum, denoted as . Instantaneous frequency offset integral , For instantaneous spectral shift, For the current moment, At the initial moment, For continuous phase after untangling, and These represent the initial position equivalent transmission length and inherent phase offset of the traveling cable, respectively. This integral characteristic is only related to the relative displacement of the car. It is directly proportional to the running speed and is independent of the running speed. When stationary, it retains the current displacement information and does not return to zero.

[0097] Offline calibration establishes calibration equations for the aforementioned parameters and the current equivalent transmission length S and bending radius R of the traveling cable. This is used to verify the physical model and provide prior constraints for the data-driven elevator shaft feature-height mapping table. The offline calibration establishes these calibration equations for the aforementioned parameters and the current equivalent transmission length S and bending radius R of the traveling cable, thereby enabling the online phase to be determined by the measured... , , The current cable configuration is deduced in reverse. Each time the system is powered on and initialized, the current four-dimensional feature vector is read, and the closest height is found in the elevator shaft feature-height mapping table as the initial position. The standard ψ value at that height is set as the initial value for the current integral. Thereafter, ψ is accumulated according to the above integral method to ensure the continuity of the displacement integral after a power outage and restart.

[0098] Finally, the winding phase difference, reflection spectrum amplitude, and instantaneous frequency offset integral are used to construct the three-dimensional cable feature. The radio frequency characteristic RSSI is periodically scanned by the 4G / 5G DTU module and used as an auxiliary verification source. Its update rate (1~10Hz) is lower than that of the cable feature update rate (20~50Hz). The most recently available value or linear interpolation is asynchronously fused with the cable feature to jointly construct a four-dimensional position feature vector. Before feature comparison, each dimension is normalized and weighted according to its importance, giving higher weight to the electromagnetic features of the cable and lower weight to RSSI as an auxiliary verification item, to ensure that the main positioning source plays a dominant role in the similarity measurement.

[0099] In some implementations, the above-mentioned 3) establishing the elevator shaft feature-height mapping table and adaptive maintenance mechanism can also be achieved in the following ways: 3.1) Establish a unified physical coordinate system, the process of which is as follows: The first-level truth value (offline) uses a laser rangefinder or a high-precision rotary encoder to establish the initial mapping table; the second-level truth value (online) uses a leveling magnetic switch or a rotary encoder for EWMA incremental updates. A laser rangefinder is used as a temporary offline calibration tool, with the first-level truth value as the truth reference source. The elevator car is stopped at the lowest level, and this point is defined as the physical height zero point.

[0100] The elevator travels at a constant low speed (e.g., 0.2 m / s to 0.5 m / s) from the bottom floor to the top floor. During this process, whenever the laser rangefinder detects a 1 cm change in the car's displacement, it outputs a hardware trigger pulse to the FPGA. Upon receiving the pulse, the FPGA immediately reads and records the current FFT result, thus achieving synchronous trigger acquisition accurate to 1 cm.

[0101] As the car rises, high-frequency signals are continuously acquired and calculated, including the winding phase difference, reflection spectrum amplitude, instantaneous frequency offset integral, and RSSI four-dimensional position feature vector. To overcome random noise, multiple samplings are performed at each height sampling point. Median filtering is used to remove obvious outliers from the M sets of raw data, and then the average value is calculated to generate the standard X vector for that height.

[0102] Because the accompanying cable may experience drastic impedance changes or excessive signal attenuation at certain characteristic locations, leading to data loss or unreliability at some sampling points, curve fitting and spline interpolation are performed. The cleaned discrete standard X vector is arranged in height order, and continuous curves are plotted for the three dimensions of winding phase difference, reflection spectrum amplitude, and instantaneous frequency offset integral, varying with height. For individual invalid data points, a cubic spline interpolation algorithm is used to fit and complete the data based on the adjacent valid data points. The cubic spline interpolation employs natural boundary conditions (i.e., the second derivative is zero at both ends) to ensure a smooth transition of the interpolation curve at the boundaries and avoid introducing spurious oscillations.

[0103] Finally, a mapping relationship is established between each standard X vector and its corresponding physical height, and an elevator shaft feature-height mapping table is established, with some data shown in Table 1.

[0104] Table 1. Elevator Shaft Characteristics-Height Mapping Table

[0105] 3.2) Online adaptive maintenance phase: An adaptive mapping table update mechanism is established. This stage is executed online during daily elevator operation and is a background maintenance mechanism for the mapping table, running in parallel with the real-time positioning process. In subsequent daily elevator operation, each time the car reaches a leveling position, the secondary truth value (leveling magnetic switch or rotary encoder) is used as the truth feedback to calculate the deviation between the current measured feature vector and the standard vector at the corresponding height in the mapping table. Simultaneously, a truth value confidence assessment is added: if the deviation between the secondary truth value and the current DTW coarse positioning result is greater than 5cm, the truth value is deemed unreliable, and the update is rejected. Further quantitative statistical criteria are added to distinguish deviation types: the overall root mean square of the deviation vectors for all leveling positions (one per floor) is calculated. and the root mean square of the difference between the deviation vectors of adjacent layers. ,like >Preset threshold and If this is determined to be a global systemic deviation (cable replacement), the adaptive recalibration mechanism for cable replacement will be activated; if >Preset threshold and If the deviation between the measured feature vector of the current floor level and the standard vector at the corresponding height in the elevator shaft feature-height mapping table exceeds a preset threshold (e.g., Euclidean distance greater than 0.05), then a normal online learning update is triggered. For normal online learning updates, the current measured feature vector is first compensated for temperature and humidity to match the calibration environment. In the formula, This is the current feature vector after temperature and humidity compensation; This is the currently measured feature vector; The temperature drift coefficient represents the change in the eigenvector caused by a 1°C change in temperature. The current temperature; For calibration temperature; The current relative humidity is expressed in %RH. To calibrate relative humidity; The humidity drift coefficient represents the change in the feature vector caused by a 1% change in relative humidity. The compensated features are then fused with the old values ​​of the elevator shaft feature-height mapping table using the exponentially weighted moving average (EWMA) algorithm. In the formula, For the updated standard vector, The old standard vector before the update. This is the current feature vector after temperature and humidity compensation. The smoothing factor is 0.1 to 0.3.

[0106] Subsequently, environmental data was collected using existing elevator environmental sensors and temperature and humidity sensors at the top, middle, and bottom of the shaft. A segmented compensation mechanism based on the shaft temperature and humidity gradient model was established: the shaft was divided into three segments along the height direction—upper, middle, and lower. Each segment used the nearest temperature and humidity sensor data for local temperature and humidity joint drift compensation, ensuring that the standard vector in the elevator shaft feature-height mapping table always corresponds to the calibrated environment. Temperature and humidity drift coefficient. and The calibration method is as follows: During the elevator installation and commissioning phase, a combined temperature and humidity working condition matrix is ​​simulated using external temperature control and humidification equipment. Four-dimensional feature vectors are collected at each temperature point under working conditions of -10℃, 10℃, 30℃, and 50℃, and relative humidity of 30%RH, 60%RH, and 90%RH, respectively. A binary linear regression model is then established. In the formula, For temperature Predicted eigenvector values ​​under relative humidity (RH) conditions; The feature vector under the calibration environment; For temperature variables; For calibration temperature; The variable is relative humidity. The result was obtained through least squares fitting. and Finally, the average of multiple measurements is taken as the final drift coefficient. Furthermore, an adaptive recalibration mechanism for cable replacement is established, introducing multi-source diagnostic decisions to distinguish cable replacement from other disturbances: when the deviation between the measured feature vectors at multiple consecutive leveling locations and the standard vector at the corresponding height in the elevator shaft feature-height mapping table exceeds a preset threshold, the existing environmental sensor data and signal injection module output power are read first; if the temperature and humidity exceed the calibration compensation range, joint temperature and humidity drift compensation is performed first, and then it is determined whether the deviation still exceeds the limit after compensation; if the injection power is abnormal, the injection module fault is investigated first; if the aforementioned quantitative statistical criteria have determined it to be a cable replacement event and the temperature, humidity, and injection power are all normal, the corresponding recalibration process is initiated; if it has been determined to be a loose mechanical constraint point, only the mapping table correction for that local area is triggered. If the cable is replaced with the same specification and model, the rapid recalibration mode is activated, and the mapping table is incrementally corrected within 2 to 3 complete operating cycles using the feedback of the most recent M leveling true values; if the cable is replaced with a different specification, model or material, the full-stroke recalibration process is triggered to re-establish the elevator shaft feature-height mapping table.

[0107] In some implementations, the online operation phase described in 4) above can also be achieved in the following ways: During the online detection phase, the broadband radio frequency receiving front-end at the car end receives high-frequency signals based on the aforementioned broadband direct sampling of the radio frequency layer and variable sampling rate triggering mechanism of the feature layer. After preprocessing, it calculates the current four-dimensional position feature vector X in real time. Subsequently, this vector is uploaded to the processing unit through the 4G / 5G DTU module as input to the matching algorithm.

[0108] Because of minute random fluctuations in the electromagnetic environment, and the mechanical vibrations of cables also introduce noise, it is not possible to simply determine whether they are equal; similarity calculation is required. This embodiment introduces the Sakoe-ChibaBand DTW algorithm as a sequence similarity metric, limiting the search range to a coarse localization neighborhood (e.g., ±50cm) to reduce computational load. Specifically, the four-dimensional position feature vector X (N ranges from 10 to 50, adaptively adjusted according to the elevator's current speed, with larger values ​​for low speeds and smaller values ​​for high speeds) is extracted from the N most recent consecutive moments to form a feature sequence. In the feature-height mapping table, each candidate sequence is truncated within this neighborhood using the same window length. And calculate its DTW distance. ,in For the monotonic alignment path between Q and C, The distance is the four-dimensional Euclidean distance. The database subsequence that minimizes the DTW distance is searched, and its corresponding height interval is the coarse positioning result of the current car. Subsequently, within this height neighborhood, the DTW matching result is refined in the online stage using lookup table interpolation (such as nearest neighbor interpolation or linear interpolation). Using the physical model (i.e., the constructed segmented leaky waveguide model and cable tension-impedance coupling equation) as prior constraints, post-processing refinement is performed using Physical-Informed Gaussian Process Regression (PI-GPR), outputting the car height accurate to the centimeter level. This height value can be further mapped to the corresponding floor.

[0109] The above description is merely an embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and scope of the present invention are included within the scope of protection of the present invention.

[0110] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of the invention. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of the invention, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the invention. The sequence numbers of the above-described embodiments of the invention are merely descriptive and do not represent the superiority or inferiority of the embodiments.

[0111] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, or apparatus. Without further limitations, an element defined by the phrase "comprising a..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. In the several embodiments provided by this invention, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components may be combined, or integrated into another system, or some features may be ignored or not performed.

[0112] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for elevator position detection based on the impedance characteristics of a traveling cable, wherein the traveling cable forms an absolute position anchor point related to the car height at a fixed mechanical constraint point due to impedance mismatch; characterized in that, The method includes: A high-frequency sweep signal is injected into the traveling cable at the elevator machine room end, and the full-band reflection signal of the high-frequency sweep signal is received in real time at the elevator car end via broadband radio frequency. Based on the processing of the full-band reflected signal, a multi-dimensional position feature vector is constructed, which includes the entanglement phase difference feature between the reflected wave and the incident wave, the reflection spectrum amplitude feature, the instantaneous frequency offset integral feature, and the wireless channel feature. The output rate of the multidimensional position feature vector is dynamically adjusted according to the real-time running speed of the car, and the multidimensional position feature vectors of the current time and multiple consecutive times before the current time are arranged in sequence to form a real-time feature sequence. An analytical mapping relationship between characteristic impedance and car height is established based on the mechanical morphology of the traveling cable, and an elevator shaft characteristic-height mapping table is established based on the analytical mapping relationship. The real-time feature sequence is matched with each candidate sequence in the elevator shaft feature-height mapping table to obtain a coarse positioning height. Interpolation refinement is performed in the neighborhood of the coarse positioning height, and post-processing refinement is performed in combination with the parsing mapping relationship as a prior constraint to output the car height value.

2. The method according to claim 1, characterized in that, The process of injecting a high-frequency sweep signal into the traveling cable at the elevator machine room end includes: An external signal coupling device is used at the equipment room end to inject a high-frequency sweep signal with a frequency range of 1MHz-1GHz into the traveling cable through broadband transformer isolation coupling or capacitor coupling; the high-frequency sweep signal avoids the communication frequency band of the elevator control system and the frequency band occupied by the operator's wireless equipment; The elevator car end utilizes a broadband radio frequency transceiver platform integrating a high-speed ADC and an FPGA to directly sample the high-frequency sweep signal at a sampling rate of no less than 2 GSPS, and the FPGA performs FFT spectrum analysis in real time to obtain the full-band reflection signal.

3. The method according to claim 1, characterized in that, The winding phase difference feature is represented as follows: In the formula, The characteristic of the entangled phase difference is that... Let be the signal propagation constant. The current equivalent transmission length of the accompanying cable. This is due to the inherent phase shift; The amplitude characteristics of the reflection spectrum are scattering parameters. frequency domain amplitude spectrum The amplitude value at the resonant frequency; The instantaneous frequency offset integral feature is expressed as: In the formula, This is the integral characteristic of instantaneous frequency offset. For instantaneous spectral shift, For integration variables, For the current moment, At the initial moment, The initial position of the accompanying cable is the equivalent transmission length. The wireless channel characteristic is the RSSI (Radio Signal Strength Indicator) of the location of the elevator car; The multidimensional position feature vector is .

4. The method according to claim 1, characterized in that, The step of dynamically adjusting the output rate of the multi-dimensional position feature vector based on the acquired real-time operating speed of the car includes: The elevator's acceleration and velocity data are acquired in real time using a six-axis inertial measurement unit integrated at the car end. With a preset spatial displacement threshold as the trigger condition, when the IMU integral displacement reaches the threshold, a feature vector is output, so that the spatial sampling interval of the elevator is constant when it runs at all speeds. The instantaneous frequency offset integral feature is obtained by integrating the continuous phase over time after phase unwrapping, and an initial phase difference is assigned based on the initial position feature vector during power failure and restart to ensure the continuity of the displacement integral.

5. The method according to claim 1, characterized in that, The establishment of an analytical mapping relationship between characteristic impedance and car height based on the mechanical configuration of the accompanying cable includes: The traveling cable is divided into a suspended section, a curved section, and a horizontal section along the hoistway height direction, which are determined by the car height. A segmented leakage waveguide model of the traveling cable is then established. Establish the cable tension-impedance coupling equation: ; In the formula, Let be the characteristic impedance of the i-th segment of the traveling cable. This is the nominal characteristic impedance of the accompanying cable. and These are the nominal inductance and nominal capacitance per unit length of the accompanying cable, respectively. The tension of the i-th segment of the traveling cable The resulting change in inductance Car height The amount of change in capacitance caused by bending deformation; The analytical mapping relationship between the characteristic impedance of each cable segment and the car height is established using the segmented leakage waveguide model and the cable tension-impedance coupling equation.

6. The method according to claim 1, characterized in that, The elevator shaft feature-height mapping table is pre-established through the following offline calibration steps: Using an external high-precision measuring device as the true height reference source, the lowest floor level position of the elevator is defined as the physical height zero point; The elevator is controlled to run from the bottom floor to the top floor at a preset uniform speed. At the same time, the car displacement is detected by the external high-precision measuring device. Whenever the car displacement reaches the preset spatial sampling interval, the multi-dimensional position feature vector corresponding to the current height is triggered to be collected and recorded. The multidimensional position feature vectors of each height sampling point are filtered and interpolated to establish an elevator shaft feature-height mapping table that corresponds one-to-one with the physical height and the multidimensional position feature vectors.

7. The method according to claim 1, characterized in that, The step of performing sequence similarity matching between the real-time feature sequence and each candidate sequence in the elevator shaft feature-height mapping table to obtain the coarse positioning height includes: In the feature-height mapping table, the preset neighborhood range of the previous positioning result is used as the search window, and each candidate sequence within the search window is truncated according to the sequence length of the real-time feature sequence; The sequence distance between the real-time feature sequence and each of the candidate sequences is calculated using a dynamic time warping algorithm. The height interval corresponding to the candidate sequence with the smallest sequence distance is selected as the coarse positioning height.

8. The method according to claim 1, characterized in that, The process involves interpolating and refining within the neighborhood of the coarse positioning height, and then combining this with the analytical mapping relationship as a priori constraints for post-processing refinement, outputting the car height value, including: Within the neighborhood corresponding to the coarse positioning height, the coarse positioning height is refined by nearest neighbor interpolation or linear interpolation to obtain a preliminary refined height. Using the segmented leaky waveguide model and the cable tension-impedance coupling equation as constraints, the preliminary refined height is post-processed and refined by physical information Gaussian process regression, and the car height value is output.

Images