Hotel safe remote centralized management and control system and method based on internet of things

Abstract

The application discloses a hotel safe remote centralized management and control system and method based on Internet of Things, and relates to the technical fields of intelligent security and Internet of Things data processing, which comprises the following steps: firstly, collecting and preprocessing the time sequence state data of the room edge perception network; secondly, constructing a trajectory fitting index by analyzing the deviation degree of the target actual point cloud trajectory and the standard intention vector; then, based on the spatial asymmetry of the personnel density in the interaction area and the waiting area, distinguishing between legal two-person cooperation and single-person overreach, and constructing a personnel distribution index; subsequently, analyzing the anti-interference characteristics of the space-time state in the global background noise to construct a space-time authorization significance index; finally, jointly evaluating the unlocking request according to the significance index, and when the value is greater than a preset safety threshold, activating a conditional trigger type dynamic token and issuing it to the safe through near field communication. The application solves the security risk that the traditional password lock is extremely easy to cause single-person overreach of the internal personnel familiar with the system.

Description

Technical Field

[0001] This invention relates to the field of intelligent security and Internet of Things (IoT) data processing technology, specifically to a remote centralized control system and method for hotel safes based on the Internet of Things. Background Technology

[0002] Traditional hotel room safes mostly operate offline, relying heavily on preset master passwords or mechanical back keys for opening in emergencies such as when guests forget their passwords. This poses a serious single point of failure security risk; once the master key is leaked, it can easily lead to unauthorized internal theft. Furthermore, traditional devices lack status awareness and network traceability capabilities, resulting in weak hotel asset security and a serious lack of credibility in post-incident accountability.

[0003] To overcome the isolated limitations of standalone devices, existing technologies generally incorporate remote centralized management solutions based on the Internet of Things (IoT). This solution constructs a centralized management network by equipping the safe with a wireless communication module and connecting it to a local area network (LAN) or the cloud. When an emergency opening is required, the centralized management platform can bypass local verification and issue a remote opening command via the network, thereby reclaiming physical access and achieving a certain degree of online monitoring of the device.

[0004] However, the above solutions still have the following problems: First, remote unlocking commands are completely detached from the physical constraints of the guest room environment. If the centralized management platform is hacked or the operating account is illegally hijacked, it is very easy to trigger a catastrophic cybersecurity incident involving a large number of abnormal unlocks. Second, the existing system lacks cross-verification in multiple modal and spatiotemporal dimensions. Relying solely on command issuance logs is insufficient to prevent single-person crimes committed by insiders familiar with the system's operating rules, making it difficult for the hotel to prove its innocence when customer complaints occur. Finally, if conventional optical video surveillance is introduced into the guest rooms to confirm the identity of the on-site operator, it will inevitably seriously infringe on the privacy of hotel guests' living spaces, leading to an irreconcilable technical conflict between security control and privacy compliance.

[0005] Therefore, the present invention provides a remote centralized management and control system and method for hotel safes based on the Internet of Things. Summary of the Invention

[0006] The purpose of this invention is to provide a remote centralized management and control system and method for hotel safes based on the Internet of Things, so as to solve the existing problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a remote centralized management and control system for hotel safes based on the Internet of Things, comprising: The multimodal environment data preprocessing module collects and preprocesses the time-series status data of the guest room edge sensing network after centralized authorization. The trajectory matching index calculation module analyzes the deviation characteristics between the actual point cloud trajectory and the standard intent vector during the movement of the target person toward the safe, and constructs the trajectory matching index. The personnel distribution index calculation module, based on the trajectory fitting index, analyzes the spatial asymmetry of personnel density in the safe interaction area and the guest room waiting area to distinguish between legitimate two-person collaboration and single-person unauthorized isolated behavior, and constructs the personnel distribution index. The spatiotemporal authorization significance index calculation module, based on the personnel distribution index, analyzes the isolated anti-interference characteristics of the legal spatiotemporal state within the effective time window in the global background noise, and constructs the spatiotemporal authorization significance index; The dynamic encryption authorization execution module jointly labels the current digital unlocking request based on the spatiotemporal authorization saliency index. When the saliency index is greater than the preset security threshold, it activates the condition-triggered dynamic token and sends it to the safe via near-field communication.

[0008] A further improvement of this invention is that the multimodal environmental data preprocessing module records the full lifecycle events of personnel activities in the guest room through the networked smart door lock and FMCW millimeter-wave radar in the guest room. This includes recording the door lock opening carrier ID through timestamps, extracting the radar intermediate frequency signal, performing distance and Doppler calculations through fast Fourier transform, and constructing a dataset of original spatial point cloud trajectory sequences with time as the independent variable and three-dimensional spatial coordinates and velocity as dependent variables.

[0009] A further improvement of this invention is that the construction process of the trajectory matching index calculation module includes: taking the moment when the door lock opens as the center, and taking the total length backward as... The time window is denoted as the monitoring window, and the spatial point cloud trajectory sequence within the monitoring window is denoted as the local motion monitoring sequence. The centroid of the local motion monitoring sequence is projected onto the horizontal plane as the input of the least squares method for path linear fitting, and the output is the fitted travel direction equation. Based on the travel direction equation, the expected value of the standard spatial coordinates under different timestamps is calculated. The difference between the actual collected coordinates and the expected value of the standard spatial coordinates is denoted as the spatial fitting residual. The sequence formed by the spatial fitting residuals is denoted as the fitting residual sequence, and then the trajectory fitting index is obtained.

[0010] A further improvement of this invention is that the personnel distribution index calculation module calculates the local motion monitoring sequence... Then, the guest room space is divided into a first subspace and a second subspace, with the midpoint of the normal vector of the plane where the safe is located as the dividing surface, representing the remote waiting area; the point cloud envelope volume boundary of the local motion monitoring sequence in the first and second subspaces is calculated by the DBSCAN clustering algorithm to obtain the personnel distribution index.

[0011] A further improvement of this invention is that the spatiotemporal authorization significance index calculation module is used to calculate the personnel distribution index for each time sliding window. Then, the sequence of personnel distribution indices from all windows is denoted as the asymmetric spatial distribution sequence. This sequence is used as the input to the automatic multi-scale peak detection algorithm, and the output is the peak point representing the stable establishment of the dual-person confirmation state. The total number of peak points obtained is denoted as... , with the first Taking the first peak point as an example, the first peak point... Centered on the peak points, with a time span of The neighborhood of is denoted as the effective authorized neighborhood, and the spatiotemporal authorization saliency index is obtained.

[0012] A further improvement of this invention is that the dynamic encryption authorization execution module performs probability mapping on the spatiotemporal authorization significance index through Bayesian inference. If the spatiotemporal authorization significance index is greater than the preset security confidence threshold, it is determined that the current guest room is in a legal multimodal spatiotemporal state. The control platform calls the ECDSA elliptic curve signature algorithm and uses the state label and timestamp to activate the signature of the preset dormant digital token to generate the downlink control payload.

[0013] On the other hand, the present invention provides a method for remote centralized management and control of hotel safes based on the Internet of Things, comprising the following steps: S1. Based on the high-level request of the centralized management and control platform, collect the carrier characteristics of the target guest room's networked smart door lock and the spatial point cloud data of millimeter-wave radar and perform filtering preprocessing; S2. Calculate and analyze the fitting residual between the target movement trajectory and the standard intention vector, and construct a trajectory fitting index that reflects the trajectory anomaly. S3. Based on the trajectory fitting index, analyze the spatial distribution characteristics to calculate the point cloud envelope volume of the safe interaction area and the waiting area, and construct a personnel distribution index to distinguish between single-person theft and double-person law enforcement. S4. Use the AMPD algorithm to extract peak values ​​of asymmetric sequences, establish an effective authorization neighborhood through time window integration, and construct a spatiotemporal authorization significance index that characterizes the stability of multimodal states. S5. Compare the spatiotemporal authorization significance index with the confidence threshold to drive the ECDSA algorithm to activate the condition-triggered dynamic token, issue an unlock command in the near field, and perform a hash operation on the data snapshot containing the above feature index and store it in the WORM storage medium.

[0014] A further improvement of this invention is that the formula for calculating the trajectory fitting index is expressed as: in, Indicates time window Trajectory fit index within the range, This represents the maximum absolute value of all Euclidean distances in the fitted residual sequence. This represents a logarithmic function with the natural constant as its base, configured to smooth out extreme value fluctuations caused by spatially dispersed distributions. This represents the standard deviation of the fitted residual sequence. This represents the mean of the absolute values ​​of all elements in the fitted residual sequence.

[0015] A further improvement of this invention is that the formula for calculating the personnel distribution index is expressed as follows: in, Indicates time window The population distribution index within the area This represents the mean of the trajectory-fitting exponential sequence. and Let represent the ranges of the envelope volumes of the personnel target point clouds in the first and second subspaces, respectively. This represents the Laplace smoothing hyperparameter.

[0016] A further improvement of this invention is that the construction process of the spatiotemporal authorization significance index is as follows: in, The spatiotemporal authorization significance index represents the k-th effective state peak point in the sequence. This represents the population distribution index at the k-th peak point. Indicates the time span of the effective authorization neighborhood. This represents the Mahalanobis distance between the s-th monitoring time and the center time of the k-th peak point within the neighborhood. Indicates the number of neighbors within the neighborhood. The multi-source living asymmetric distribution index at time 1.

[0017] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention firstly extracts the target trajectory fit index and personnel distribution index to jointly determine the physical space status of the guest room at the edge, and uses this as the sole basis for activating the near-field dynamic token; it solves the security risk that traditional password locks are prone to being opened by internal personnel familiar with the system without authorization; it realizes the binding of high-level unlocking with real time and space status, blocking single-point failure and internal crime routes.

[0018] 2. By using a multi-source cross-wake mechanism on the edge side and a read-and-burn data cleaning method that overwrites the underlying memory, the compliance conflicts and data redundancy issues caused by traditional radar all-weather scanning are resolved. Security verification can be completed without optical image acquisition and without the original trajectory leaving the guest room, thus taking into account both the security needs of hotel assets and privacy protection standards. Attached Figure Description

[0019] Figure 1 This is a framework diagram of a remote centralized control system for hotel safes based on the Internet of Things (IoT) according to the present invention. Figure 2 This is a flowchart of a remote centralized management method for hotel safes based on the Internet of Things (IoT) according to the present invention. Detailed Implementation

[0020] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations thereof. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0021] The term "and / or" is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three cases: A exists alone, A and B exist simultaneously, and B exists alone.

[0022] Example 1 This embodiment is applied to the remote centralized management and control scenario of hotel room safes, and is especially suitable for scenarios where guests forget their passwords or need to unlock the safe in case of sudden illness. It can effectively solve the technical problems of hotels being unable to prove their innocence after emergency unlocking, the single authorization chain, and the lack of credibility of audit logs.

[0023] Figure 1 This embodiment presents a framework diagram of an IoT-based remote centralized control system for hotel safes, including: a multimodal environment data preprocessing module, a trajectory matching index calculation module, a personnel distribution index calculation module, a spatiotemporal authorization saliency index calculation module, and a dynamic encrypted authorization execution module. The equipment used includes: a remote centralized control platform deployed in the hotel monitoring center, equipped with a data processing server, WORM storage media, and an ECDSA elliptic curve signature module; a guest room edge sensing network, with each guest room equipped with one networked smart lock and one FMCW millimeter-wave radar; a hotel safe with a built-in near-field communication module, dynamic token receiving module, and unlocking mechanism; and a handheld terminal for security personnel supporting physical card reading and authentication functions. All devices establish communication connections with the centralized control platform through the hotel's local area network, enabling real-time data transmission and command interaction.

[0024] Conventional standalone storage lockers typically use offline combination locks or simple card swipe mechanisms. When guests forget their passwords or in case of emergencies requiring management intervention, the standard solution is to directly open the locker using a master password or a mechanical back key. While this conventional solution can open the locker door, it inevitably introduces a single point of failure, creating a security vulnerability. If the master password is leaked or the master key is obtained by insiders, internal theft is highly likely, and it is difficult to prove innocence afterward.

[0025] To address the unauthorized access risks posed by conventional mechanical or simple electronic backdoors, the industry has introduced remote door-opening command solutions based on wide area network (WAN) connections. However, remote command issuance is not constrained by the physical environment of the guest room. If the cloud platform is hacked or the front desk terminal is illegally hijacked, attackers could easily achieve the catastrophic consequence of remotely opening all doors with a single click. Furthermore, installing optical cameras inside guest rooms to verify occupancy would severely infringe upon guests' privacy. Therefore, this application introduces a multimodal environmental perception and edge-side collaborative decision-making architecture. By deploying smart safe terminals, networked guest room door lock terminals, and frequency-modulated continuous wave millimeter-wave radar, it achieves the binding of high-privilege commands to the physical space of the guest room without acquiring any optical images.

[0026] Based on this, the system first collects and preprocesses the time-series state data of the guest room edge sensing network after centralized authorization through the multimodal environmental data preprocessing module. In conventional security monitoring, sensors are often in a state of continuous scanning around the clock. Although this conventional approach can capture the trajectory of all personnel, it generates a large amount of invalid and redundant data, and also causes guests to have privacy concerns about 24 / 7 scanning. To solve the data redundancy and privacy concerns caused by 24 / 7 scanning, this module introduces a pre-sleep and multi-source cross-wake mechanism in the edge-side microcontroller. The microcontroller remains in a suspended state by default. Only when the guest room's networked smart door lock uploads the door opening signal of the security-level RFID card, and the system simultaneously receives a temporary work order token containing a high-authority identifier from the cloud, will the microcontroller transition to an active state. After activation, the module extracts the radar intermediate frequency signal, calls the Fast Fourier Transform to convert it into distance and Doppler frequency shift parameters, and then generates a raw spatial point cloud trajectory sequence dataset with time as the independent variable and three-dimensional spatial coordinates and velocity as dependent variables.

[0027] While the aforementioned wake-up mechanism addresses privacy concerns, the raw point cloud data collected contains significant noise due to multipath effects, leading to serious false alarms in subsequent assessments if used directly. To eliminate noise interference and identify the movement intent of individuals, the system invokes the trajectory fit index calculation module. This module analyzes the deviation between the actual point cloud trajectory and the standard intent vector as the target person moves towards the safe, constructing a trajectory fit index. A conventional method for determining movement intent involves setting an absolute distance threshold; anyone entering a specific radius is considered to have intentionally approached. This conventional method completely fails to distinguish between normal straight-line walking and abnormal loitering, probing, or rummaging behavior. To address the inability to identify abnormal trajectories, this module takes the total length backward from the moment the door lock is opened as the center. The time window is denoted as the monitoring window. Since the time it typically takes for security personnel to walk from entering a guest room to reaching the temporary storage equipment is between ten and twenty seconds, the half-width parameter of the time window is... A preferred value of 15 can fully cover the core signal segment of the motion cycle. The sequence within the monitoring window is recorded as the local entity motion monitoring sequence. After the centroid is projected onto the horizontal plane, a least-squares path linear fitting is performed to output the travel direction equation and calculate the expected value of the standard spatial coordinates at different timestamps. The difference between the Euclidean distance between the actual collected coordinates and the expected value of the standard spatial coordinates is recorded as the spatial fitting residual, forming a fitting residual sequence, and thus obtaining the trajectory fitting index.

[0028] If the radar is maliciously blocked, resulting in insufficient sampling and a very small number of elements in the fitted residual sequence, the system will activate an invalid point replacement scheme, forcibly filling the sequence with a very large penalty bias value. Introducing this index allows the system to filter out irregular wandering trajectories. However, if an insider familiar with the room layout uses a smooth, straight-line trajectory to illegally activate the equipment, the above algorithm will fail, thus failing to prevent unauthorized single-person attacks. As a precaution against zero-valued positive numbers, this prevents the algorithm from crashing when the logarithm within the parentheses is less than zero when the trajectory is absolutely straight.

[0029] To address the issue that the trajectory alignment index cannot intercept skilled insiders committing crimes alone, the system utilizes a personnel distribution index calculation module. While adding a facial recognition component to confirm whether two people are traveling together is feasible, this violates the compliance guidelines for systems without optical imaging. This module, based on the trajectory alignment index, analyzes the spatial asymmetry of personnel density between the safe interaction area and the guest room waiting area to distinguish between legitimate two-person collaboration and isolated, unauthorized behavior, thus constructing a personnel distribution index. The guest room space is divided into a first subspace (representing the near-end interaction area) and a second subspace (representing the far-end waiting area), using the midpoint of the normal vector of the plane where the safe is located as the dividing surface. A density-based spatial clustering algorithm is used to calculate the point cloud envelope volume boundary within the subspace, yielding the personnel distribution index. The average projected width of an adult's torso is approximately 0.4 to 0.5 meters; in this embodiment, the neighborhood radius parameter of the clustering algorithm is preferably 0.45, which precisely encapsulates a single, independent living person into a spatial cluster.

[0030] At this point, the low-frequency vibrations generated by the cleaning cart pushing in the actual corridor can cause transient multipath interference to the radar, leading to false high values ​​in the distribution index. To solve the problem of false high values ​​caused by transient multipath interference, the system calls the spatiotemporal authorization significance index calculation module. The existing method is to apply a simple moving average filter to smooth the entire signal, but this will smooth out the originally sharp and effective action signal, resulting in a sluggish system response. This module analyzes the isolated anti-interference performance of the legal spatiotemporal state within the effective time window in the global background noise, and uses an automatic multi-scale peak detection algorithm to extract asymmetric sequence peaks. The total number of peak points obtained is denoted as . , with the first The neighborhood centered on the peak point and with a time span of D is denoted as the effective authorized neighborhood, and the spatiotemporal authorization significance index is obtained. Since completing a legal near-field handshake authentication requires at least twenty cycles of a stable spatial state, abrupt signals with a width less than this are likely vibration disturbances. Therefore, the time span parameter can preferably be 20 in this embodiment.

[0031] After obtaining a robust saliency index, the dynamic encrypted authorization execution module is run. A conventional security solution involves directly uploading action logs to a central database. However, the central database is easily tampered with by high-privilege accounts, resulting in a lack of credibility in hotel logs when customer complaints arise. To address the lack of credibility in traceability, this module uses Bayesian inference to perform a probability mapping on the spatiotemporal authorization saliency index and compares this index with a preset security confidence threshold. If the security confidence threshold is too low, it is susceptible to noise and false activations; if it approaches a natural number (1), it will lead to repeated failures of legitimate authorization. Therefore, the optimal value for the security confidence threshold parameter is 0.82, which balances usability and robustness.

[0032] If the spatiotemporal authorization significance index is determined to be greater than the security confidence threshold, the control platform invokes the elliptic curve signature algorithm, using state tags and timestamps to activate the dormant digital token, generating a short-lived downlink control payload. After verification, the microcontroller drives the unlocking process and performs a hash operation on a data snapshot containing all the aforementioned computational indices, storing it in a write-multiple-read storage medium.

[0033] Example 2 Figure 2 This embodiment illustrates a flowchart of a remote centralized management method for hotel safes based on the Internet of Things (IoT). Based on the same inventive concept as Embodiment 1, this invention provides a remote centralized management method for hotel safes based on the IoT, with the following steps: S1. Based on the high-level request of the centralized management and control platform, collect the carrier characteristics of the target guest room's networked smart door lock and the spatial point cloud data of millimeter-wave radar and perform filtering preprocessing; S2. Calculate and analyze the fitting residuals between the target movement trajectory and the standard intention vector, and construct a trajectory fitting index that reflects the trajectory anomaly. The formula for calculating the trajectory fitting index is as follows: ,in, Representing a time window Trajectory fit index within; To fit the maximum value of the absolute value of all Euclidean distances in the residual sequence, an upper limit threshold of 2.5 can be preferred. If it is greater than this maximum value, it indicates that the motion trajectory has serious disordered reversal, reflecting a clear search intention. It is a logarithmic function with the natural constant as its base, configured to smooth out extreme value fluctuations caused by spatial hashing. The standard deviation of the fitted residual sequence reflects the degree of dispersion of the trajectory deviation; To fit the mean of the absolute values ​​of all elements in the residual sequence; a constant 1 is used as a zero-prevention minimum positive number to avoid the algorithm crashing when the result of taking the logarithm within the brackets is less than zero when the trajectory is absolutely straight.

[0034] S3. Based on the trajectory fitting index, analyze the spatial distribution characteristics to calculate the point cloud envelope volume of the safe interaction area and the waiting area, and construct a personnel distribution index to distinguish between single-person theft and two-person law enforcement; the formula for the personnel distribution index is expressed as: ,in, For time window, The population distribution index within the area The mean of the trajectory-fitting exponential sequence. and These represent the range of the envelope volume of the personnel target point cloud in the first subspace and the second subspace, respectively, which directly reflects the fullness of the entity occupancy in the corresponding space; This represents a hyperparameter; its function is to prevent calculation anomalies where the denominator is zero when no personnel are detected occupying space in the second subspace.

[0035] When there is one and only one person in the first subspace Approaching zero results in a minimal final output; however, when meeting compliant two-person collaboration requirements, It has a significantly non-zero value, thus outputting a high-level signal.

[0036] S4. The AMPD algorithm is used to extract peak values ​​from asymmetric sequences. An effective grant neighborhood is established through time window integration, and a spatiotemporal grant significance index characterizing the stability of multimodal states is constructed. The formula for calculating the spatiotemporal grant significance index is as follows: ,in, The spatiotemporal authorization significance index represents the k-th effective state peak point in the sequence. This represents the population distribution index at the k-th peak point. Indicates the time span of the effective authorization neighborhood. This represents the Mahalanobis distance between the s-th monitoring time and the center time of the k-th peak point within the neighborhood, reflecting the feature drift over time. Indicates the number of neighbors within the neighborhood. The multi-source live asymmetric distribution index at each time step. If the peak value represents genuine two-person collaboration, the integral term increases significantly; if it represents transient perturbation, the integral term decreases sharply, filtering out false alarms.

[0037] S5. Compare the spatiotemporal authorization significance index with the confidence threshold to drive the ECDSA algorithm to activate the condition-triggered dynamic token, issue an unlock command in the near field, and perform a hash operation on the data snapshot containing the above feature index and store it in the WORM storage medium.

[0038] The threshold and weight settings involved in this embodiment can be set by default according to the present invention, or can be set by those skilled in the art.

[0039] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0040] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0041] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0042] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0043] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A remote centralized management and control system for hotel safes based on the Internet of Things, characterized in that: The system includes: The multimodal environment data preprocessing module collects and preprocesses the time-series status data of the guest room edge sensing network after centralized authorization. The trajectory matching index calculation module analyzes the deviation characteristics between the actual point cloud trajectory and the standard intent vector during the movement of the target person toward the safe, and constructs the trajectory matching index. The personnel distribution index calculation module, based on the trajectory fitting index, analyzes the spatial asymmetry of personnel density in the safe interaction area and the guest room waiting area to distinguish between legitimate two-person collaboration and single-person unauthorized isolated behavior, and constructs the personnel distribution index. The spatiotemporal authorization significance index calculation module, based on the personnel distribution index, analyzes the isolated anti-interference characteristics of the legal spatiotemporal state within the effective time window in the global background noise, and constructs the spatiotemporal authorization significance index; The dynamic encryption authorization execution module jointly labels the current digital unlocking request based on the spatiotemporal authorization saliency index. When the saliency index is greater than the preset security threshold, it activates the condition-triggered dynamic token and sends it to the safe via near-field communication.

2. The IoT-based remote centralized control system for hotel safes according to claim 1, characterized in that: The multimodal environmental data preprocessing module records the entire lifecycle of human activities in guest rooms through networked smart door locks and FMCW millimeter-wave radar. This includes recording the door lock's opening carrier ID with timestamps, extracting the radar's intermediate frequency signal, performing distance and Doppler calculations through fast Fourier transform, and constructing a dataset of original spatial point cloud trajectory sequences with time as the independent variable and three-dimensional spatial coordinates and velocity as dependent variables.

3. The IoT-based remote centralized control system for hotel safes according to claim 2, characterized in that: The construction process of the trajectory matching index calculation module includes: taking the moment when the door lock opens as the center, and taking the total length backward as... The time window is denoted as the monitoring window, and the spatial point cloud trajectory sequence within the monitoring window is denoted as the local motion monitoring sequence. The centroid of the local motion monitoring sequence is projected onto the horizontal plane as the input of the least squares method for path linear fitting, and the output is the fitted travel direction equation. Based on the travel direction equation, the expected value of the standard spatial coordinates under different timestamps is calculated. The difference between the actual collected coordinates and the expected value of the standard spatial coordinates is denoted as the spatial fitting residual. The sequence formed by the spatial fitting residuals is denoted as the fitting residual sequence, and then the trajectory fitting index is obtained.

4. The IoT-based remote centralized control system for hotel safes according to claim 3, characterized in that: The personnel distribution index calculation module calculates the local motion monitoring sequence. Then, the guest room space is divided into a first subspace and a second subspace, with the midpoint of the normal vector of the plane where the safe is located as the dividing surface, representing the remote waiting area; the point cloud envelope volume boundary of the local motion monitoring sequence in the first and second subspaces is calculated by the DBSCAN clustering algorithm to obtain the personnel distribution index.

5. The IoT-based remote centralized control system for hotel safes according to claim 4, characterized in that: The spatiotemporal authorization significance index calculation module is used to calculate the personnel distribution index for each time sliding window. Then, the sequence of personnel distribution indices from all windows is denoted as the asymmetric spatial distribution sequence. This sequence is used as the input to the automatic multi-scale peak detection algorithm, and the output is the peak point representing the stable establishment of the dual-person confirmation state. The total number of peak points obtained is denoted as... , with the first Taking the first peak point as an example, the first peak point... Centered on the peak points, with a time span of The neighborhood of is denoted as the effective authorized neighborhood, and the spatiotemporal authorization significance index is obtained.

6. The IoT-based remote centralized control system for hotel safes according to claim 5, characterized in that: The dynamic encryption authorization execution module performs probability mapping on the spatiotemporal authorization significance index through Bayesian inference. If the spatiotemporal authorization significance index is greater than the preset security confidence threshold, it is determined that the current guest room is in a legal multimodal spatiotemporal state. The control platform calls the ECDSA elliptic curve signature algorithm and uses the state label and timestamp to activate the signature of the preset dormant digital token to generate the downlink control payload.

7. A method for remote centralized control of hotel safes based on the Internet of Things (IoT), used to execute the remote centralized control system for hotel safes based on the IoT as described in any one of claims 1-6, characterized in that: Includes the following steps: S1. Based on the high-level request of the centralized management and control platform, collect the carrier characteristics of the target guest room's networked smart door lock and the spatial point cloud data of millimeter-wave radar and perform filtering preprocessing; S2. Calculate and analyze the fitting residual between the target movement trajectory and the standard intention vector, and construct a trajectory fitting index that reflects the trajectory anomaly. S3. Based on the trajectory fitting index, analyze the spatial distribution characteristics to calculate the point cloud envelope volume of the safe interaction area and the waiting area, and construct a personnel distribution index to distinguish between single-person theft and double-person law enforcement. S4. Use the AMPD algorithm to extract peak values ​​of asymmetric sequences, establish an effective authorization neighborhood through time window integration, and construct a spatiotemporal authorization significance index that characterizes the stability of multimodal states. S5. Compare the spatiotemporal authorization significance index with the confidence threshold to drive the ECDSA algorithm to activate the condition-triggered dynamic token, issue an unlock command in the near field, and perform a hash operation on the data snapshot containing the above feature index and store it in the WORM storage medium.

8. The method for remote centralized management and control of hotel safes based on the Internet of Things according to claim 7, characterized in that: The formula for calculating the trajectory matching index is as follows: in, Indicates time window Trajectory fit index within the range, This represents the maximum absolute value of all Euclidean distances in the fitted residual sequence. This represents a logarithmic function with the natural constant as its base, configured to smooth out extreme value fluctuations caused by spatially dispersed distributions. This represents the standard deviation of the fitted residual sequence. This represents the mean of the absolute values ​​of all elements in the fitted residual sequence.

9. The method for remote centralized management and control of hotel safes based on the Internet of Things according to claim 8, characterized in that: The formula for calculating the personnel distribution index is as follows: in, Indicates time window The population distribution index within the area This represents the mean of the trajectory-fitting exponential sequence. and Let represent the ranges of the envelope volumes of the personnel target point clouds in the first and second subspaces, respectively. This represents the Laplace smoothing hyperparameter.

10. The method for remote centralized management and control of hotel safes based on the Internet of Things according to claim 9, characterized in that: The construction process of the spatiotemporal authorization significance index is as follows: in, The spatiotemporal authorization significance index represents the k-th effective state peak point in the sequence. This represents the population distribution index at the k-th peak point. Indicates the time span of the effective authorization neighborhood. This represents the Mahalanobis distance between the s-th monitoring time and the center time of the k-th peak point within the neighborhood. Indicates the number of neighbors within the neighborhood. The multi-source living asymmetric distribution index at time 1.

Images