A computer host protection system, case

Abstract

The present application relates to the field of detection, in particular to a computer host protection system and a case. The system comprises an isolation unit, a sensing unit and a protection unit. The isolation unit includes a backup power supply and an interlocking switch. The sensing unit is provided with a magnetic induction device and a vibration coil to monitor the closed loop communication state of the case structure, collect electromagnetic signals and establish electromagnetic fingerprints, calculate the evolution slope and related characteristic parameters, build a situation space and generate warning information. The protection unit obtains the warning information, adjusts the excitation current of the device, and completes the trajectory comparison and verification. The present application forms a structure monitoring closed loop by arranging a magnetic induction device, constructs an electromagnetic fingerprint by a vibration coil, extracts features to build a situation space to identify external objects, adjusts the excitation current to form interference and verifies the trajectory, relies on an independent backup power supply to realize active real-time protection at all times, solves the problems of traditional protection lag, blind area, electromagnetic eavesdropping and the like, and effectively safeguards the data confidentiality and hardware integrity of the host.

Description

Technical Field

[0001] This invention relates to the field of information protection, and in particular to a computer host protection system and chassis. Background Technology

[0002] With the deepening of information technology infrastructure, the value of computer host data continues to increase, making its security a core aspect of the complex web of network and physical security. Highly sensitive scenarios require defense against both network logic attacks and physical hardware intrusions, necessitating the construction of multi-layered protection. Traditional host protection primarily relies on software encryption and access control, depending on the host's CPU and operating system, using software instructions to manage hardware and achieve resource protection. This model is highly dependent on the host's power-on and operational status; when powered off or in standby mode, monitoring blind spots emerge, making it impossible to detect physical disassembly and hardware modifications.

[0003] To address the blind spots in host power outage protection, some technologies have attempted to add independent hardware to implement power outage early warning. For example, the power outage chassis intrusion detection system disclosed in Chinese patent application CN115238265A relies on the CMOS battery short-circuit mechanism to detect power outages and unpacking. If the server is opened after a power outage, the chassis intrusion device triggers a CMOS battery short circuit, resulting in CMOS data loss. After the system is powered back on, the BIOS detects that the system time has returned to the factory default value and immediately reports the intrusion information to the BMC via IPMI commands. The BMC then sends notifications to the maintenance personnel via web page prompts and alarm emails. This solution is cost-effective and effectively compensates for the blind spot in server protection where the BMC cannot detect chassis intrusions during a power outage, ensuring the security of server hardware and data.

[0004] However, the aforementioned power-off chassis intrusion detection scheme is a passive protection mechanism. It only triggers a warning message via BIOS detection after the chassis has been opened and the host has been powered on again. Subsequent protection measures still rely on manual intervention by maintenance personnel after receiving the warning message. External attackers can use a miniature data acquisition device to detect electromagnetic signals emitted by the chassis and steal keys before the power is cut off. Maintenance personnel cannot detect the opening of the chassis in time and make adequate protective preparations. After the host is powered on again, external attackers can use the obtained key to quickly steal information while maintenance personnel are receiving and processing the warning message, resulting in a very high risk of data theft for the host. Summary of the Invention

[0005] This invention provides a computer host protection system and chassis to solve the problem of high data theft risk associated with passive early warning host power failure protection methods.

[0006] To solve the above-mentioned technical problems, this application provides the following technical solution: A computer host protection system, comprising: The isolation unit includes a backup power supply independent of the host and an interlock switch located at the host physical interface; The sensing unit includes magnetic induction devices deployed at various locations within the chassis. These magnetic induction devices are connected in series via a signal loop at the chassis connection gaps, and the signal loop is connected to the backup power supply to form a structural monitoring closed loop. It also includes a resonant coil for emitting electromagnetic signals. The sensing unit is used to monitor the connectivity of the structural monitoring closed loop. It collects real-time electromagnetic signals at various locations within the chassis through the magnetic induction devices, extracts the electromagnetic fingerprint reference by statically distributing the real-time electromagnetic signals, and obtains the real-time evolution slope value from the temporal changes of the real-time electromagnetic signals. The protection unit is used to obtain the connectivity status of the structural monitoring closed loop, and send the location of the structural monitoring closed loop in the chassis as protection information to the operation and maintenance terminal based on the connectivity status. When sending the protection information, the host physical interface is closed through the interlock switch. The sensing unit is also used to acquire the change characteristics of the real-time evolution slope value and extract continuous fluctuation characteristics; based on the continuous fluctuation characteristics, it acquires the impedance characteristics and energy attenuation shadow characteristics of each monitoring loop, and constructs a situation space by combining the impedance characteristics and energy attenuation shadow characteristics, and acquires the volume, velocity and position of external objects as warning information from the situation space; the protection unit acquires the risk value based on each warning information, and sends the warning information as protection information to the operation and maintenance terminal based on the risk value. The protection unit is also used to adjust the excitation current of each magnetic induction device according to the risk value and the preset current sequence to obtain the interference magnetic field, acquire the changes of warning information in the interference magnetic field, and extract the motion trajectory of the external object; when adjusting the excitation current, it sends the verification information to the operation and maintenance terminal and receives the displacement verification trajectory. Based on the difference between the displacement verification trajectory and the motion trajectory, the warning information is sent to the operation and maintenance terminal as protection information.

[0007] The basic principle and beneficial effects of this invention are as follows: This invention deploys magnetic induction devices at various locations within the chassis and connects them in series at the joints to form a structural monitoring closed loop. This loop, combined with a resonant coil emitting electromagnetic signals and collecting real-time electromagnetic signals, extracts features from the real-time electromagnetic signals when the chassis is intact and not subject to electromagnetic theft. Based on the extraction results, an electromagnetic fingerprint baseline is constructed. This baseline is then compared with the real-time electromagnetic signals during chassis operation to obtain the real-time evolution slope value. Based on the slope features, the impedance characteristics and energy attenuation shadow characteristics of the monitoring loop are extracted, and a situational awareness space is constructed. Changes in the situational awareness space accurately identify relevant parameters of external objects. Simultaneously, the excitation current of the magnetic induction devices is adjusted according to the risk value to create an interference magnetic field, completing the extraction and verification of the external object's trajectory. Continuous operation is ensured by an independent backup power supply. When chassis deformation or high-risk intrusion is detected, the physical interface is disconnected, and protection information is pushed. This achieves proactive real-time monitoring and protection, promptly blocking illegal intrusion and data theft from both physical structure and electromagnetic environment dimensions, ensuring the security of data and hardware within the host.

[0008] This invention adjusts the excitation current of the magnetic induction device to create an interfering magnetic field, which can effectively disrupt the transmission and analysis process of electromagnetic signals when external objects approach, thereby preventing the illegal theft and analysis of electromagnetic signals inside the host. Simultaneously, it combines the verification trajectory feedback from maintenance personnel to complete trajectory comparison and security screening of external objects, eliminating suspicious objects and significantly improving the accuracy and reliability of defense judgments. This protection method combining interference and verification is suitable for use scenarios where classified equipment is under constant monitoring. When an unidentified detection device approaches the chassis attempting to steal electromagnetic signals, the interfering magnetic field can directly disrupt its signal acquisition process, while trajectory verification can quickly distinguish between normal inspections by maintenance personnel and illegal detection activities, avoiding the triggering of protective actions due to misjudgment and affecting normal use. It can also identify and interrupt illegal theft activities when unattended.

[0009] In this invention, a closed-loop monitoring system formed by a series of magnetic induction devices works in conjunction with an electromagnetic fingerprint reference to create a dual verification mechanism. This mechanism accurately verifies the integrity of the chassis structure, quickly pinpoints chassis deformation or opening, and leverages dynamically changing electromagnetic signals to increase the difficulty of signal analysis by external stealing devices. This makes it difficult for external devices to quickly separate the actual host electromagnetic signals, thus strengthening signal theft prevention capabilities at the signal level. This dual verification mechanism is suitable for scenarios where the host is forcibly disassembled after power failure. Traditional protection methods cannot monitor the chassis status during power failure, but this invention can promptly detect disassembly and block data leakage through closed-loop switching and electromagnetic fingerprint changes. Even if attackers attempt to steal electromagnetic signals before disassembly, they will be unable to obtain effective information due to interference from dynamic electromagnetic signals, thereby solving the problem of hardware and data being easily stolen during power failure.

[0010] This invention combines situational awareness construction with external object parameter recognition to predict intrusion risks before the chassis structure deforms. It shifts the defense focus from the post-intrusion stage to the intrusion prediction stage, achieving proactive early warning and pre-emptive protection, effectively avoiding the lag issues of traditional protection methods. This pre-emptive prediction capability is suitable for scenarios where attackers gradually approach the chassis, preparing to dismantle it. When an external object slowly approaches the chassis, this invention can reconstruct the object's volume, speed, and position through impedance characteristics and energy attenuation shadows, predicting in advance whether it will damage the chassis structure. It eliminates the need to wait for chassis deformation to activate protection, proactively disconnecting physical interfaces and sending warning messages in high-risk situations, allowing protection actions to precede intrusion and avoiding the passive situation of remediation after data theft has occurred.

[0011] This invention relies on an independent backup power supply for monitoring and protection, allowing for continuous operation independent of the host's power supply limitations. It provides round-the-clock security monitoring even when the host is not in operation, filling the security gap during non-operational periods. Simultaneously, it uses the movement trajectory of external objects as both a trigger for proactive defense and a reference for security verification, leveraging sensor data to improve defense execution efficiency and reduce unnecessary protective operations caused by misjudgments. This round-the-clock monitoring and efficient data utilization feature is suitable for unattended office or server room scenarios after the host is powered off. Even when the host is powered off, this invention can continuously monitor the chassis structure and surrounding electromagnetic environment, ensuring protection capabilities are not lost due to host shutdown. Furthermore, the dual application of trajectory data accurately identifies intrusion behavior, making the overall protection process more rational and practical, comprehensively safeguarding host hardware and data security.

[0012] In summary, this invention forms a closed-loop structural monitoring system by deploying magnetic induction devices, constructs an electromagnetic fingerprint using resonant coils, extracts features to construct a situational awareness space to identify external objects, adjusts the excitation current to generate interference and verify the trajectory, and relies on an independent backup power supply to achieve active real-time protection at all times. This solves the problems of traditional protection lag, blind spots, and electromagnetic theft, and effectively ensures the confidentiality of host data and the integrity of hardware.

[0013] Furthermore, the magnetic induction device is embedded in the side wall of the chassis and arranged in an array; the sensing unit performs cross-point feature correlation and linkage fitting on the electromagnetic signals at multiple points according to the embedding position of the magnetic induction device on the chassis, integrates them to form the global electromagnetic features of the chassis, and constructs an electromagnetic fingerprint reference based on the global electromagnetic features of the chassis; the sensing unit obtains the inherent attenuation law of the electromagnetic signal corresponding to the connection gap according to the positional relationship between each electromagnetic signal, and corrects the global electromagnetic features of the chassis according to the inherent attenuation law during the construction of the electromagnetic fingerprint reference.

[0014] This invention integrates the electromagnetic characteristics of the entire chassis by embedding magnetic induction devices into the sidewall of the chassis and arranging them in an array. It combines cross-point feature correlation and linkage fitting to correct the electromagnetic fingerprint reference based on the inherent law of electromagnetic signal attenuation at connection gaps, thereby improving the accuracy of the reference calibration. This not only reduces the impact of single-point electromagnetic signal interference but also makes the electromagnetic fingerprint reference more closely match the actual chassis structure. In scenarios involving classified host operation, it can effectively resist electromagnetic signal interference from external miniature acquisition devices, prevent key theft by induction, and ensure the accuracy of subsequent electromagnetic signal detection. Simultaneously, it lays a reliable foundation for obtaining subsequent impedance characteristics and energy attenuation shadow characteristics.

[0015] Furthermore, the chassis forms a closed cavity structure and covers and fixes the host inside the closed cavity. The resonant coil is fixedly connected to the inner wall of the chassis cavity. When no current is connected to the resonant coil, the sensing unit collects electromagnetic signals as the host electromagnetic radiation signals. The sensing unit combines the assembly structure formed between the host and the chassis with the host electromagnetic signals, marks the electromagnetic radiation concentration area on the inner wall of the chassis, and determines the setting position of the resonant coil on the chassis based on the electromagnetic radiation concentration area.

[0016] This invention secures the host computer within a closed enclosure, with a resonant coil fixed to the inner wall of the enclosure. The coil's position is determined based on the concentration area of ​​electromagnetic radiation from the host computer, achieving both stable control of the host and optimized electromagnetic radiation coverage by the resonant coil. This not only enhances the overall protection of the enclosure but also allows the resonant coil to match the host's radiation characteristics. For example, during standby or low-power operation, it can specifically generate a protective magnetic field, reducing electromagnetic leakage and preventing attackers from stealing data through induced electromagnetic signals. Simultaneously, it reduces the resonant coil's energy consumption and improves protection efficiency.

[0017] Furthermore, based on the inherent attenuation law at the connection gap, the sensing unit obtains the features corresponding to the preset reference state of the electromagnetic fingerprint as the reference features, and obtains the features corresponding to the reference state in the real-time electromagnetic signal as the control features. Combining the acquisition time of the real-time electromagnetic signal and the difference between the reference features and the control features, time-series fitting is performed to obtain the real-time evolution slope value. The preset reference state includes the spatial distribution offset state corresponding to the spatial field distribution of the electromagnetic signal, the characteristic mode fluctuation state corresponding to the time-series waveform of the alternating electromagnetic signal, and the inter-array linkage feature offset state corresponding to the signal coupling of the array points of the chassis gap series signal loop array.

[0018] This invention combines the inherent attenuation law of connection gaps, compares benchmark features with control features, and performs time-series fitting to obtain real-time evolution slope values. Multiple reference states are preset to ensure the comprehensiveness and accuracy of the slope values. This not only provides reliable data support for subsequent acquisition of impedance characteristics and energy attenuation shadow characteristics, but also captures subtle temporal changes in electromagnetic signals. For example, when an attacker approaches the chassis or electromagnetic signals exhibit weak fluctuations, it can accurately identify signal change trends, avoiding detection deviations caused by signal fluctuations. The principle lies in covering different signal change scenarios through multiple reference states, thereby improving the robustness of data extraction.

[0019] Furthermore, the sensing unit combines the baseline features, comparison features, and continuous fluctuation features corresponding to the preset reference state to compare the amplitude changes and phase shifts of the real-time electromagnetic signals of each monitoring loop with the baseline features, and then obtains the impedance features of each monitoring loop by combining the acquisition time; it combines the inherent attenuation baseline of the global electromagnetic features, the inherent law of electromagnetic signal attenuation at the connection gaps, and the law of electromagnetic signal energy loss change extracted from the continuous fluctuation features of the real-time evolution slope value to obtain the propagation loss distribution of the electromagnetic signal in and around the preset warning range of the chassis, and maps the propagation loss distribution to form the energy attenuation shadow feature; the sensing unit performs multi-dimensional fusion of the amplitude and phase parameters of the impedance features with the loss distribution and attenuation amplitude dimension of the energy attenuation shadow feature, calibrates the spatial coordinates corresponding to the impedance features and the energy attenuation shadow feature, and constructs a situation space covering the preset warning range of the chassis; by comparing the differences between the impedance features, energy attenuation shadow features and the baseline features in the situation space, it selects areas within the preset warning range where both the impedance shift and energy attenuation are greater than the preset abnormal threshold as abnormal areas. The sensing unit obtains the volume of the external object based on the magnitude of the impedance shift in the abnormal region in the situation space, combined with the coverage area of ​​the energy attenuation shadow feature in the situation space, and according to the preset binary feature calibration relationship; it tracks the spatial coordinate changes of the impedance and energy attenuation features corresponding to the abnormal region in the time-series continuous fluctuation feature, and obtains the speed of the external object's movement by combining the interval of the real-time electromagnetic signal acquisition time; and it locates the position of the external object within the preset warning range based on the situation space coordinates corresponding to the abnormal region.

[0020] This invention constructs a situational awareness space by fusing multidimensional impedance features and energy attenuation shadow features, filters out abnormal regions, and acquires external object parameters to achieve accurate identification and location of external objects. It can not only comprehensively capture anomalies around the chassis, but also obtain the volume and velocity of external objects through binary feature calibration relationships. For example, when an attacker slowly approaches the chassis with a miniature acquisition device, it can be accurately identified through changes in energy attenuation shadows. By using shadow features formed by propagation loss distribution mapping, it can capture changes in the electromagnetic environment after the object's intervention, preventing malicious approach from going undetected.

[0021] Furthermore, the sensing unit sends warning information to the protection unit. The protection unit randomly generates an interference trajectory based on the position and speed of the external object in the warning information. The risk value corresponding to any point on the interference trajectory is less than a preset collision risk value. According to the relationship that the excitation current amplitude, operating frequency and spatial distance are inversely proportional, and the excitation current timing trigger interval is directly proportional to the spatial distance of the external object, the excitation current is adjusted according to the interference trajectory. The protection unit combines the dynamic changes of the excitation current with the electromagnetic fingerprint reference to extract the electromagnetic fingerprint reference corresponding to the interference trajectory and construct an electromagnetic interference fingerprint. The protection unit compares the electromagnetic interference fingerprint with the multi-dimensional feature differences in the situation space and obtains the motion trajectory of the external object based on the feature differences.

[0022] This invention extracts the trajectory of external objects by randomly generating interference trajectories and adjusting the excitation current, combined with electromagnetic interference fingerprinting and situational space comparison. This approach avoids the risk of collisions with external objects while improving trajectory extraction accuracy. It not only dynamically adapts to the motion state of external objects but also ensures trajectory accuracy through verification information. For example, when an attacker rapidly approaches the chassis with a data acquisition device, the excitation current can be adjusted to create an interference magnetic field to track their trajectory, while simultaneously preventing the interference magnetic field from affecting the normal operation of the host. The principle lies in offsetting the influence of environmental electromagnetic interference by comparing electromagnetic interference fingerprints with situational space features.

[0023] Furthermore, the sensing unit collects the amplitude of the signal circuit impedance characteristics at the chassis connection gap and the corresponding time-series change period; the sensing unit extracts the vibration amplitude between the chassis sidewalls at the chassis gap from the impedance characteristic amplitude, and extracts the vibration period of the chassis sidewalls at the chassis gap from the time-series change period of the impedance characteristic amplitude; the protection unit obtains the inherent correlation between the impedance value at the chassis gap and the gap width from the electromagnetic fingerprint reference, establishes an impedance gap mapping relationship, substitutes the real-time collected impedance value at the chassis gap into the impedance gap mapping relationship, and obtains the real-time gap width at the chassis connection gap; the protection unit adjusts the current amplitude and current phase of the corresponding signal circuit at the chassis gap according to the expansion or contraction trend of the real-time gap width; the protection unit controls the current phase, current phase duration, and switching sequence of the magnetic induction device on the corresponding signal circuit according to the vibration period. When the chassis gap width shows an expansion trend, it controls adjacent magnetic induction devices to form an electromagnetic attraction effect; when the chassis gap width shows a contraction trend, it controls adjacent magnetic induction devices to form an electromagnetic repulsion effect.

[0024] This invention extracts vibration parameters by collecting the impedance characteristics of the gap, establishes an impedance-gap mapping relationship to obtain the real-time gap width, and adjusts the electromagnetic attraction and repulsion based on the vibration period and the gap change trend to effectively buffer gap deformation. It not only protects the chassis structure but also adapts its protective actions to vibration conditions. For example, when the fan is running or the chassis is accidentally bumped, electromagnetic attraction and repulsion can counteract the vibration impact, preventing gap expansion and electromagnetic leakage. By leveraging the linkage between electromagnetic attraction and repulsion forces, vibration period, and gap changes, dynamic protection is achieved, reducing structural damage.

[0025] Furthermore, the sensing unit obtains the real-time gap width corresponding to each connection gap in the chassis based on the impedance gap mapping relationship, establishes a spatial coordinate system with the chassis geometric center as the origin, and obtains the geometric position of each connection gap in the spatial coordinate system as the gap position based on the chassis geometric dimensions sent by the maintenance terminal. The gap position is then associated with and stored in relation to the real-time gap width. When the protection unit receives the placement and locking command sent by the maintenance terminal, it solidifies and stores the real-time gap width, gap position, and movement reference data, using the coordinates of the origin of the spatial coordinate system as the initial centroid coordinates. The protection unit monitors the numerical changes in the real-time gap width and adjusts the initial centroid coordinates in conjunction with the gap position to obtain the moving centroid coordinates. The system combines the acquisition time of the real-time gap width to obtain the moving center of gravity coordinates in the spatial coordinate system, the moving orientation and speed of the moving center of gravity coordinates, the distance between the moving center of gravity coordinates and the initial center of gravity coordinates as the offset distance, and extracts the vibration verification period from the moving orientation and speed. When the difference between the vibration verification period and the vibration period is less than the preset difference threshold and the offset distance is less than the preset stability threshold, the current amplitude and current phase of the signal circuit are adjusted according to the changing trend of the real-time gap width. The protection unit is also used to send the preset moving signal as protection information to the operation and maintenance terminal when the offset distance is not less than the preset offset stability threshold or the difference between each adjacent vibration verification period is greater than the preset vibration stability threshold.

[0026] This invention establishes a spatial coordinate system, calculates the center of gravity coordinates and offset parameters, and combines vibration verification cycle and offset distance to determine the chassis status, achieving accurate differentiation between normal vibration and malicious handling. It not only avoids false alarms from normal vibration but also clearly detects malicious handling. For example, when a worker accidentally touches the chassis during a routine inspection, causing slight vibration, a false alarm can be ruled out through parameter comparison. Conversely, when an attacker maliciously moves the chassis to disassemble it, protection information can be quickly sent, and gap data can be stored simultaneously to support subsequent traceability, improving the accuracy and reliability of the protection.

[0027] Furthermore, the sensing unit is also used to collect the real-time temperature inside the chassis, and to perform amplitude compensation on the real-time electromagnetic signal according to the preset temperature sensitivity correction curve. The amplitude-compensated real-time electromagnetic signal is used for static distribution extraction and real-time evolution slope value acquisition. When the difference between the vibration verification period and the vibration period is less than the preset difference threshold, and the offset distance is less than the preset stability threshold, the protection unit combines the combination of vibration period and vibration amplitude with the corresponding real-time temperature to establish a temperature vibration correlation. Before adjusting the current amplitude and current phase of the signal circuit according to the changing trend of the real-time gap width, the combination of the current vibration period and vibration amplitude is verified through the temperature vibration correlation. Based on the verification result, the preset moving signal is sent to the operation and maintenance terminal as protection information.

[0028] This invention collects real-time internal temperature data from the chassis, compensates for electromagnetic signal amplitude based on a temperature sensitivity correction curve, and verifies the rationality of vibration by combining temperature and vibration correlation. This improves vibration detection accuracy and eliminates false alarms caused by normal vibration. It not only corrects the influence of temperature on electromagnetic signals but also accurately distinguishes between normal vibration and malicious movement. For example, when the chassis vibrates due to the operation of the built-in cooling module, false alarms can be eliminated through temperature correlation verification. Furthermore, it accurately identifies and sends protection information when an attacker maliciously shakes the chassis to test it. By utilizing the correlation between temperature and normal vibration, it filters out non-malicious vibration signals, reducing the false alarm rate. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the functional modules of a computer host protection system. Figure 2 This is a structural schematic diagram of the chassis and the magnetic induction device on the side wall. Figure 3 A timing diagram illustrating the relationship between impedance characteristic acquisition and gap width calculation; Explanation of reference numerals in the attached diagram: 1. Chassis; 2. Magnetic induction device. Detailed Implementation

[0030] The following will describe the concept and technical effects of the present invention clearly and completely with reference to embodiments, so as to fully understand the purpose, features and effects of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are all within the scope of protection of the present invention. Example 1 like Figure 1As shown, a computer host protection system (hereinafter referred to as the "protection system") includes an isolation unit, a sensing unit, and a protection unit. The protection unit and part of the sensing unit are integrated into a control chip independent of the host, and the control chip is located inside the chassis 1 and engaged with the chassis 1. The control chip can be a general-purpose industrial control chip.

[0031] The maintenance terminal adopts an industrial-grade touch-screen terminal, equipped with conventional wired Ethernet, wireless, and low-power wireless communication modules. It is used to receive protection information and verification information sent by the protection system, and to issue placement and locking commands to the protection system. The maintenance terminal is used by maintenance personnel to view information and issue operation commands.

[0032] The chassis 1 is constructed from multiple metal plates, forming a closed cavity structure that encloses and secures the main unit within it. Chassis 1 is made of conventional metal with good electromagnetic shielding properties; in this embodiment, sheet metal is used. Sheet metal is a conventional metal material with good electromagnetic compatibility and structural rigidity, and will not cause shielding interference to the electromagnetic signal acquisition of the magnetic induction device 2. The main unit is fixed inside the closed cavity of chassis 1 using conventional fixing methods (such as nut fixing). A reasonable heat dissipation gap is provided between the main unit casing and the inner wall of chassis 1 to prevent deformation and damage to chassis 1 due to temperature changes. The magnetic induction device 2 forms a signal loop in series at this gap, which can clearly monitor the opening and closing state of the gap and changes in surrounding electromagnetic signals.

[0033] The isolation unit includes a backup power supply independent of the host and an interlock switch located at the host's physical interface. The backup power supply uses a conventional energy storage battery, which is electrically connected to the power supply interface of the control chip, providing power support for the normal operation of the protection system after the host is powered off. The interlock switch is a conventional electromagnetic interlock switch, which is electrically connected to the control chip, and the control chip outputs commands to control the switch's on / off state.

[0034] The sensing unit includes magnetic induction devices 2 deployed at various locations on the chassis 1. The magnetic induction devices 2 employ conventional Hall effect sensors, such as... Figure 2 As shown, the magnetic induction device 2 is embedded in the side wall of the chassis 1 and arranged in an array. The magnetic induction device 2 is connected in series through a signal circuit at the connection gap of the chassis 1. The signal circuit is connected to the backup power supply to form a closed loop for structural monitoring. The maintenance personnel preset the conduction resistance of the signal circuit through the maintenance terminal. When the opening and closing state of the gap changes, the resistance of the signal circuit will change synchronously with the gap width. The sensing unit monitors the connection status of the structural monitoring closed loop by monitoring the change of the circuit resistance.

[0035] The sensing unit also includes a resonant coil for emitting electromagnetic signals. When no current is connected to the resonant coil, the sensing unit collects real-time electromagnetic signals from various locations on the chassis 1 through the arrayed magnetic induction devices 2. The unit sets the acquisition frequency and duration, and uses a Gaussian filtering algorithm to extract the static distribution of the collected real-time electromagnetic signals, obtaining an electromagnetic fingerprint baseline. Simultaneously, it obtains the real-time evolution slope value from the temporal changes of the real-time electromagnetic signals. Specifically, based on the embedding position of the magnetic induction devices 2 on the chassis 1, the sensing unit performs cross-point feature correlation and linkage fitting on the multi-point electromagnetic signals (using the least squares method, with the amplitude of each point's electromagnetic signal as the dependent variable and the acquisition time as the independent variable, fitting a linear equation; the slope of the equation is the real-time evolution slope value), integrating these to form the global electromagnetic features of the chassis 1, and constructing an electromagnetic fingerprint baseline based on these global electromagnetic features.

[0036] Maintenance personnel pre-enter the control chip with the geometric dimensions and the positions of each magnetic induction device 2 on the chassis 1 (the geometric dimensions of chassis 1 and the positions of each magnetic induction device 2 on chassis 1 are sent to the control chip via the maintenance terminal). The sensing unit establishes a spatial coordinate system with the geometric center of chassis 1 as the origin, creating a three-dimensional spatial coordinate system and marking the coordinates of each magnetic induction device 2. The sensing unit obtains the inherent attenuation law of the electromagnetic signal corresponding to the connection gap based on the positional relationship between each electromagnetic signal (maintenance personnel collect electromagnetic signals through the maintenance terminal when chassis 1 is in normal working condition, and obtain the fixed rate attenuation of the electromagnetic signal by observing the change in the amplitude of the electromagnetic signal at the gap; this attenuation rate is the inherent attenuation law). During the construction of the electromagnetic fingerprint reference, the sensing unit corrects the global electromagnetic characteristics of chassis 1 according to the inherent attenuation law to avoid the impact of electromagnetic signal attenuation at the gap on the calibration accuracy of the electromagnetic fingerprint reference.

[0037] The resonant coil is fixedly engaged with the inner wall of the chassis 1 cavity. When no current is applied to the resonant coil, the sensing unit collects electromagnetic signals as the host's electromagnetic radiation signals. The sensing unit combines the assembly structure between the host and chassis 1 (i.e., the fixing method between the host and the inner wall of chassis 1, and the dimensions of the heat dissipation gap) with the host's electromagnetic radiation signals. Using a grayscale threshold segmentation algorithm, it marks the electromagnetic radiation concentration areas (i.e., areas where the electromagnetic radiation amplitude reaches the preset standard set by the maintenance personnel) on the inner wall of chassis 1. Based on these concentration areas, it determines the placement of the resonant coil on chassis 1, ensuring that the electromagnetic signals emitted by the resonant coil more completely cover these concentration areas, thus improving electromagnetic protection effectiveness. Specifically, the sensing unit receives the host's electromagnetic radiation signals and the assembly structure of chassis 1 sent by the maintenance terminal. Using the grayscale threshold set by the maintenance personnel, it performs grayscale filtering on the electromagnetic signals, marking the areas on the inner wall of chassis 1 where the radiation amplitude meets the standard as electromagnetic radiation concentration areas, thereby determining the placement of the resonant coil.

[0038] Based on the inherent attenuation law at the connection gap, the sensing unit acquires the features corresponding to the preset reference state of the electromagnetic fingerprint (i.e., the state when chassis 1 is working normally and there are no foreign objects near chassis 1) as the reference features, and acquires the features corresponding to the reference state in the real-time electromagnetic signal as the control features. Combining the acquisition time of the real-time electromagnetic signal and the difference between the reference and control features, a linear regression algorithm is used in conjunction with the fitting time set by the maintenance personnel to perform time-series fitting and obtain the real-time evolution slope value. Among them, the preset reference state includes the spatial distribution offset state corresponding to the spatial field distribution of the electromagnetic signal (the maintenance personnel preset a reasonable offset threshold, and exceeding the offset threshold is considered abnormal), the characteristic mode fluctuation state corresponding to the time-series waveform of the alternating electromagnetic signal (the maintenance personnel preset a reasonable fluctuation amplitude threshold, and exceeding the fluctuation amplitude threshold is considered abnormal), and the inter-array linkage feature offset state corresponding to the signal coupling of the array points of the series signal loop array at the gap of chassis 1 (the maintenance personnel preset a reasonable linkage offset threshold, and exceeding the linkage offset threshold is considered abnormal). Specifically, the sensing unit combines the baseline features, comparison features, and acquisition time with the fitting time set by the operation and maintenance personnel to perform time-series fitting. It fits a linear relationship with feature differences as variables and time as independent variables, solves the equation slope, and obtains the real-time evolution slope value.

[0039] The sensing unit also acquires the changing characteristics of the real-time evolution slope value and uses a sliding window algorithm (the size of the sliding window is set by the maintenance personnel) to extract the fluctuation range of the slope value within the sliding window as a continuous fluctuation feature. Based on the continuous fluctuation feature, the impedance characteristics and energy attenuation shadow characteristics of each monitoring loop are obtained. Specifically, the sensing unit combines the baseline characteristics, control characteristics, and continuous fluctuation features corresponding to a preset reference state, compares the amplitude changes and phase shifts of the real-time electromagnetic signals of each monitoring loop with the baseline characteristics, and then combines this with the acquisition time to obtain the impedance characteristics of each monitoring loop. More specifically, if there is a reasonable amplitude difference and phase shift between the baseline characteristics and the control characteristics, the impedance characteristics are calculated through the correlation between the two.

[0040] The sensing unit combines the inherent attenuation benchmark of the global electromagnetic characteristics, the inherent attenuation law of the electromagnetic signal at the connection gap, and the electromagnetic signal energy loss change law extracted from the continuous fluctuation characteristics of the real-time evolution slope value to obtain the propagation loss distribution of the electromagnetic signal within the preset warning range of chassis 1 (the preset warning range is an area extending outward from the outer wall of chassis 1 at a reasonable distance, which is an area set in advance by maintenance personnel, such as an area within 1m of chassis 1) and its surroundings. This propagation loss distribution is then mapped to form an energy attenuation shadow feature. Specifically, the propagation loss value is directly proportional to the gray value of the shadow feature; that is, the greater the propagation loss, the higher the gray value of the shadow feature. If the propagation loss reaches different levels, the corresponding gray value of the shadow feature shows a gradient change.

[0041] The sensing unit performs multi-dimensional fusion of the amplitude and phase parameters of the impedance characteristics with the loss distribution and attenuation amplitude dimensions of the energy attenuation shadow characteristics. Principal component analysis is used during multi-dimensional fusion to extract core features of reasonable dimensions for fusion. Specifically, the sensing unit performs redundant removal and feature condensation integration on the multi-dimensional parameters of the impedance and energy attenuation shadow characteristics through feature dimensionality reduction fusion, completing the multi-dimensional feature fusion to obtain the comprehensive features required for constructing the situation space. The sensing unit calibrates the spatial coordinates corresponding to the impedance and energy attenuation shadow characteristics (based on the three-dimensional spatial coordinate system established earlier, combined with the position coordinates of the magnetic induction device 2), constructing a situation space covering the preset warning range of chassis 1. By comparing the differences between the impedance and energy attenuation shadow characteristics in the situation space and the baseline characteristics, areas within the preset warning range where both impedance offset and energy attenuation are greater than the preset abnormal threshold are selected as abnormal areas.

[0042] The sensing unit obtains the volume of an external object based on the magnitude of impedance shift within an abnormal region in the situation space, combined with the coverage area of ​​energy attenuation shadow features in the situation space, and according to a preset binary feature calibration relationship. The preset binary feature calibration relationship states that the product of the impedance shift magnitude and the coverage area is directly proportional to the volume of the external object. The sensing unit tracks the spatial coordinate changes of the impedance and energy attenuation features corresponding to the abnormal region in the continuous temporal fluctuation characteristics. Combined with the time interval of real-time electromagnetic signal acquisition, it obtains the speed of the external object's movement (speed = distance between two acquisitions ÷ acquisition interval; if the distance between two acquisitions and the acquisition time interval satisfy the actual ratio between the spatial coordinate change corresponding to the abnormal region acquired by the sensing unit in real-time and the preset real-time electromagnetic signal acquisition time interval, the corresponding movement speed is obtained). Based on the situation space coordinates corresponding to the abnormal region, the position of the external object within a preset warning range is located.

[0043] The protection unit is used to acquire the connectivity status of the structural monitoring closed loop. Based on the connectivity status, it sends the location of the structural monitoring closed loop within chassis 1 as protection information to the maintenance terminal. Simultaneously, the protection unit sends a shutdown command to the interlock switch. Upon receiving the shutdown command, the interlock switch disconnects all physical interfaces of the host computer to prevent data leakage.

[0044] The protection unit is also used to calculate the probability of a collision between an intruder and chassis 1 based on a risk value. This risk value is calculated by the protection unit using trajectory extrapolation, based on the real-time position and velocity changes of an intruder entering a preset warning range, combined with the intruder's volume parameters. When the risk value reaches or exceeds the risk threshold set by the maintenance personnel, protection information is immediately sent to the maintenance terminal. When the risk value is below the threshold, the observation period is delayed by a preset duration, during which electromagnetic signals are continuously monitored, and protection information is sent to the maintenance terminal based on the detection results.

[0045] The protection unit is also used to adjust the excitation current of each magnetic induction device 2 according to the risk value and a preset current timing sequence to obtain the interference magnetic field. Specifically, when adjusting the excitation current, the amplitude and operating frequency of the excitation current are inversely proportional to the spatial distance of the external object, and the timing trigger interval of the excitation current is directly proportional to the spatial distance of the external object (if the spatial distance of the external object changes, the amplitude, operating frequency, and timing trigger interval of the excitation current are adjusted accordingly). The protection unit obtains changes in the warning information in the interference magnetic field and extracts the motion trajectory of the external object; when adjusting the excitation current, the protection unit sends verification information (including the current excitation current parameters and the intensity of the interference magnetic field) to the maintenance terminal through the communication module of the control chip. After receiving the verification information, the maintenance terminal randomly generates a displacement verification trajectory within the preset warning range that will not collide with the chassis 1, based on the current distance between itself and the chassis 1. That is, the risk value corresponding to any point on the interference trajectory is less than the preset collision risk value, and feeds it back to the protection unit. The protection unit adjusts the excitation current according to the relationship that the amplitude and operating frequency of the excitation current are inversely proportional to the spatial distance and the timing trigger interval of the excitation current is directly proportional to the spatial distance of the external object, based on the interference trajectory.

[0046] The protection unit combines the dynamic changes of the excitation current with an electromagnetic fingerprint benchmark to extract the electromagnetic fingerprint benchmark corresponding to the interference trajectory and construct an electromagnetic interference fingerprint (using a feature extraction algorithm to extract the amplitude and phase features corresponding to the excitation current changes and fuse them with the electromagnetic fingerprint benchmark). The protection unit compares the electromagnetic interference fingerprint with the multi-dimensional feature differences in the situation space (using a cosine similarity algorithm to calculate the similarity between the two, setting a reasonable similarity threshold; when the similarity is below the threshold, it is determined that external interference exists), and obtains the motion trajectory of the external object based on the feature differences to ensure the accuracy of trajectory extraction. Specifically, the protection unit uses a feature deconstruction extraction method to decompose and extract the amplitude and phase features of the dynamic change data of the excitation current and the electromagnetic fingerprint benchmark, and then fuses and reconstructs them to obtain an electromagnetic interference fingerprint for difference comparison.

[0047] Example 2 The only difference between this embodiment and Embodiment 1 is that, Figure 3 As shown, the sensing unit collects the amplitude of the signal circuit impedance characteristics and the corresponding time-series change period at the connection gap of chassis 1. The sensing unit adopts a feature correlation comparison processing method, using the impedance characteristic amplitude and time-series change period of the gap signal circuit as the basic data, and performs correlation matching comparison with the inherent parameters of the side wall gap deformation of chassis 1 and the time-series fluctuation law of electromagnetic signals respectively. Through parameter correspondence fitting analysis, the correlation between amplitude and deformation amplitude, and the correlation between period and vibration rhythm are analyzed, and finally the vibration amplitude and vibration period corresponding to the side wall at the gap position of chassis 1 are calculated.

[0048] The protection unit retrieves the impedance reference parameters and gap reference width parameters of each gap in the electromagnetic fingerprint reference under normal factory conditions of chassis 1, and establishes an impedance gap mapping relationship with the gap number as the associated dimension. The impedance gap mapping relationship is limited to the impedance value collected in real time for a single gap loop as the only input variable and the actual width of the corresponding gap as the only output variable, forming a fixed one-to-one matching relationship. The protection unit substitutes the impedance values ​​of each gap collected in real time into the impedance gap mapping relationship one by one to match and obtain the real-time gap width at each connection gap of chassis 1. The protection unit identifies the dynamic expansion or contraction trend of the real-time gap width and adjusts the current amplitude and current phase parameters of the corresponding signal loop according to the trend change.

[0049] The sensing unit uses the three-dimensional spatial coordinate system in Embodiment 1. Based on the impedance gap mapping relationship, it calculates the real-time gap width of all connection gaps in the chassis 1 in batches. It retrieves the spatial geometric coordinates of each gap that have been pre-issued and solidified by the maintenance terminal as the gap position. It binds and associates each set of gap position parameters with the corresponding real-time gap width parameters and stores them uniformly in the built-in storage area of ​​the control chip.

[0050] The protection unit uses the vibration cycle of the side wall of the chassis 1 extracted by the sensing unit as the basis for adjustment. According to the preset timing matching rules, it sets the current phase switching node, current duration and working state switching interval of each magnetic induction device 2. When the width of the gap in the chassis 1 is detected to be expanding, the adjacent two sets of magnetic induction devices 2 are adjusted to form an electromagnetic attraction. When the width of the gap in the chassis 1 is detected to be contracting, the adjacent two sets of magnetic induction devices 2 are adjusted to form an electromagnetic repulsion.

[0051] After receiving the placement and locking command from the operation and maintenance terminal, the protection unit immediately locks and stores all real-time gap widths collected at the current moment, the associated and bound gap position parameters, and the movement reference data preset by the operation and maintenance personnel for the protection system. At the same time, it calibrates the origin coordinates of the three-dimensional spatial coordinate system as the initial centroid coordinates, which serve as a fixed reference for subsequent centroid offset comparisons.

[0052] The protection unit captures the dynamic changes in the width of each connection gap in real time. Combining the spatial geometric position parameters corresponding to each gap, it uses a coordinate correction and conversion method to offset the preset initial center of gravity coordinates and solves for the moving center of gravity coordinates under real-time operating conditions. Combining the acquisition timestamps of each set of gap width data, it analyzes the offset direction of the moving center of gravity coordinates in the spatial coordinate system through time-series trajectory analysis to determine the moving direction and analyzes the continuous change law of the coordinates over time to determine the moving speed. By comparing spatial coordinate differences, it calculates the spatial interval distance between the moving center of gravity coordinates and the initial center of gravity coordinates, which is defined as the offset distance. Then, it decomposes the moving direction and moving speed into time-series features and extracts the time-series fluctuation law to form the vibration verification cycle.

[0053] The maintenance terminal sends the difference threshold and stability threshold to the control chip in advance and stores them. When the characteristic difference between the vibration verification cycle and the original vibration cycle is detected to fall within the preset difference threshold limit set by the maintenance personnel, and the center of gravity offset distance does not exceed the preset stability threshold limit, the protection unit synchronously and adaptively adjusts the current amplitude and current phase of the corresponding signal circuit according to the expansion or contraction trend of the real-time gap width.

[0054] When the detected center of gravity offset distance exceeds the preset offset stability threshold, or the fluctuation difference between multiple sets of vibration verification cycles extracted from adjacent time series exceeds the preset vibration stability threshold, the protection unit immediately generates a preset movement signal, integrates and encapsulates the movement signal into standard protection information, and sends it to the operation and maintenance terminal through an encrypted communication link to complete the alarm.

[0055] In applications involving unattended, power-off, and static protection of classified mainframes, this embodiment addresses the shortcomings of traditional protection methods that only provide passive alarms after the fact and cannot predict when the chassis 1 is pried open or maliciously moved. It leverages the physical characteristic that the deformation of the chassis 1's gaps synchronously changes the amplitude and timing of the signal circuit impedance, establishing a one-to-one mapping relationship between impedance and gap width. By combining three-dimensional spatial coordinates with the deformation parameters of each gap, it calculates the chassis 1's center of gravity offset, movement posture, and vibration rhythm. Based on the vibration cycle and gap deformation trend, it dynamically adjusts the magnetic induction device 2 to create an electromagnetic attraction and repulsion buffer. Simultaneously, it uses multi-threshold linkage verification to distinguish between normal environmental disturbances and malicious human-caused movement. This not only provides dynamic electromagnetic constraint buffering for minor deformations in the chassis 1's gaps, reducing structural losses and preventing electromagnetic signal leakage, but also accurately identifies abnormal movement of the chassis 1 and filters false alarms from non-malicious vibrations. It achieves proactive sensing, real-time judgment, and immediate alarming under power-off conditions, overcoming the shortcomings of existing protection mechanisms that are lagging behind and lack predictive capabilities.

[0056] Example 3 The difference between this embodiment and embodiments 1-2 is that a temperature sensor is embedded inside the chassis 1 near the core heat-generating area of ​​the host and near the fan. The sensing unit collects the real-time temperature inside the chassis 1 through the temperature sensor. The collection frequency is consistent with the electromagnetic signal collection frequency to ensure that the temperature data and electromagnetic signal data are synchronized in time. The collected real-time temperature data is transmitted to the core processing module of the sensing unit in real time.

[0057] The sensing unit retrieves the temperature sensitivity correction curve pre-sent and stored in the maintenance terminal. Specifically, before chassis 1 leaves the factory, maintenance personnel simulate different temperature environments (covering the normal operation range of chassis 1 and the range of ambient temperature fluctuations), collect deviation data between the standard electromagnetic signal amplitude and the actual collected amplitude at each temperature, and use a linear fitting algorithm to fit the relationship curve between temperature and amplitude deviation as the independent variable. This curve is the temperature sensitivity correction curve. The temperature sensitivity correction curve is used to compensate for the deviation of electromagnetic signal amplitude at different temperatures and eliminate the interference of temperature changes on the accuracy of electromagnetic signal acquisition.

[0058] The sensing unit substitutes the real-time temperature data collected by the temperature sensor into the temperature sensitivity correction curve, queries the amplitude deviation value corresponding to that temperature, and uses the calculation method of "compensated amplitude = real-time collected amplitude + corresponding temperature deviation value" to calibrate and compensate the amplitude of the real-time electromagnetic signal (if the amplitude of the real-time collected electromagnetic signal is a certain value, and the deviation value corresponding to the real-time temperature is a certain corresponding value, the compensated amplitude is the sum of the two); the amplitude-compensated real-time electromagnetic signal is used as the sole input data for subsequent static distribution extraction and real-time evolution slope value acquisition to ensure the accuracy of subsequent feature extraction and vibration parameter extraction.

[0059] When the protection unit detects that the difference between the vibration verification period and the original vibration period extracted by the sensing unit is less than the preset difference threshold set by the maintenance personnel, and the center of gravity offset distance is less than the preset stability threshold set by the maintenance personnel, the protection unit retrieves the vibration parameter combination (vibration period, vibration amplitude) at the current moment, and simultaneously retrieves the corresponding real-time temperature data collected by the sensing unit. Using a feature association fitting method, the temperature data and vibration parameters are correlated one-to-one to establish a temperature-vibration correlation relationship. The temperature-vibration correlation relationship is used to limit the range of normal vibration period and normal vibration amplitude corresponding to different temperature ranges. That is, the input variable is the real-time temperature, and the output variable is the normal vibration period and vibration amplitude range corresponding to that temperature, forming a fixed temperature and vibration parameter mapping rule (if a certain temperature range corresponds to a certain range of normal vibration period and a certain range of vibration amplitude, exceeding that range is judged as abnormal).

[0060] Before adjusting the current amplitude and phase of the corresponding signal circuit based on the real-time gap width change trend, the protection unit first performs vibration rationality verification. Specifically, the protection unit retrieves the vibration period and vibration amplitude combination data at the current moment, and simultaneously retrieves the real-time temperature data at the corresponding moment, substituting the three into the established temperature-vibration correlation for verification. The verification criteria are as follows: if the current vibration period and vibration amplitude both fall within the normal vibration parameter range corresponding to the real-time temperature set by the maintenance personnel, it is determined to be normal vibration (such as vibration caused by the operation of the cooling module built into chassis 1), and the current parameter adjustment operation continues to achieve buffer protection against gap deformation; if either the current vibration period or vibration amplitude exceeds the normal vibration parameter range corresponding to the real-time temperature, it is determined to be abnormal vibration (such as vibration caused by malicious shaking of chassis 1), and a preset moving signal is immediately generated. The moving signal is integrated and encapsulated into standard protection information and sent to the maintenance terminal through an encrypted communication link to complete the alarm. At the same time, the current parameter adjustment operation is suspended to avoid misjudging normal vibration and omitting malicious moving.

[0061] The above are merely embodiments of the present invention. The invention is not limited to the fields covered by these embodiments. Commonly known structures and characteristics in the solutions are not described in detail here. Those skilled in the art are aware of all common technical knowledge in the field prior to the application date or priority date, are able to access all existing technologies in that field, and have the ability to apply conventional experimental methods prior to that date. Those skilled in the art can, under the guidance of this application, improve and implement this solution in combination with their own capabilities. Some typical known structures or methods should not be obstacles for those skilled in the art to implement this application. It should be noted that those skilled in the art can make several modifications and improvements without departing from the structure of the present invention. These should also be considered within the scope of protection of the present invention, and will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A computer host protection system, characterized in that, include: The isolation unit includes a backup power supply independent of the host and an interlock switch located at the host physical interface; The sensing unit includes magnetic induction devices deployed at various locations within the chassis. These magnetic induction devices are connected in series via a signal loop at the chassis connection gaps, and the signal loop is connected to the backup power supply to form a structural monitoring closed loop. It also includes a resonant coil for emitting electromagnetic signals. The sensing unit is used to monitor the connectivity of the structural monitoring closed loop. It collects real-time electromagnetic signals at various locations within the chassis through the magnetic induction devices, extracts the electromagnetic fingerprint reference by statically distributing the real-time electromagnetic signals, and obtains the real-time evolution slope value from the temporal changes of the real-time electromagnetic signals. The protection unit is used to obtain the connectivity status of the structural monitoring closed loop, and send the location of the structural monitoring closed loop in the chassis as protection information to the operation and maintenance terminal based on the connectivity status. When sending the protection information, the host physical interface is closed through the interlock switch. The sensing unit is also used to acquire the change characteristics of the real-time evolution slope value and extract continuous fluctuation characteristics; Based on the continuous fluctuation characteristics, the impedance characteristics and energy attenuation shadow characteristics of each monitoring loop are obtained. The situation space is constructed by combining the impedance characteristics and energy attenuation shadow characteristics. The volume, velocity and position of external objects are obtained from the situation space as warning information. The protection unit obtains the risk value based on each warning message, and sends the warning message as protection information to the operation and maintenance terminal based on the risk value; The protection unit is also used to adjust the excitation current of each magnetic induction device according to the risk value and the preset current sequence to obtain the interference magnetic field, acquire the changes of warning information in the interference magnetic field, and extract the motion trajectory of the external object; when adjusting the excitation current, it sends the verification information to the operation and maintenance terminal and receives the displacement verification trajectory. Based on the difference between the displacement verification trajectory and the motion trajectory, the warning information is sent to the operation and maintenance terminal as protection information.

2. The computer host protection system according to claim 1, characterized in that: The magnetic induction device is embedded in the side wall of the chassis and arranged in an array. The sensing unit performs cross-point feature correlation and linkage fitting on the electromagnetic signals of multiple points according to the embedding position of the magnetic induction device on the chassis, integrates them to form the global electromagnetic features of the chassis, and constructs an electromagnetic fingerprint reference based on the global electromagnetic features of the chassis. The sensing unit obtains the inherent law of attenuation of the electromagnetic signal corresponding to the connection gap according to the positional relationship between each electromagnetic signal, and corrects the global electromagnetic features of the chassis according to the inherent law of attenuation during the construction of the electromagnetic fingerprint reference.

3. The computer host protection system according to claim 1, characterized in that: The chassis forms a closed cavity structure and covers and fixes the host inside the closed cavity. The resonant coil is fixedly connected to the inner wall of the chassis cavity. When no current is connected to the resonant coil, the sensing unit collects electromagnetic signals as the host electromagnetic radiation signals. The sensing unit combines the assembly structure formed between the host and the chassis with the host electromagnetic signals, marks the electromagnetic radiation concentration area on the inner wall of the chassis, and determines the setting position of the resonant coil on the chassis based on the electromagnetic radiation concentration area.

4. A computer host protection system according to claim 2, characterized in that: Based on the inherent attenuation law at the connection gap, the sensing unit obtains the features corresponding to the preset reference state of the electromagnetic fingerprint as the reference features, and obtains the features corresponding to the reference state in the real-time electromagnetic signal as the control features. Combining the acquisition time of the real-time electromagnetic signal and the difference between the reference features and the control features, the unit performs time-series fitting to obtain the real-time evolution slope value. The preset reference states include the spatial distribution offset state corresponding to the spatial field distribution of electromagnetic signals, the characteristic mode fluctuation state corresponding to the time sequence waveform of alternating electromagnetic signals, and the inter-array point linkage characteristic offset state corresponding to the signal coupling of the array points in the chassis gap series signal circuit.

5. A computer host protection system according to claim 4, characterized in that: The sensing unit combines the benchmark features, comparison features, and continuous fluctuation features corresponding to the preset reference state to compare the amplitude changes and phase shifts of the real-time electromagnetic signals of each monitoring loop with the benchmark features. It then combines the acquisition time to obtain the impedance features of each monitoring loop. The unit combines the inherent attenuation benchmark of the global electromagnetic features, the inherent attenuation law of the electromagnetic signal at the connection gap, and the electromagnetic signal energy loss change law extracted from the continuous fluctuation features of the real-time evolution slope value to obtain the propagation loss distribution of the electromagnetic signal in the preset warning range and surrounding area of ​​the chassis. The propagation loss distribution is then mapped to form the energy attenuation shadow feature. The sensing unit performs multi-dimensional fusion of the amplitude and phase parameters of the impedance characteristics with the loss distribution and attenuation amplitude dimension of the energy attenuation shadow characteristics, calibrates the spatial coordinates corresponding to the impedance characteristics and energy attenuation shadow characteristics, and constructs a situation space covering the preset warning range of the chassis. By comparing the differences between the impedance characteristics, energy attenuation shadow characteristics and the reference characteristics in the situation space, the area within the preset warning range where the impedance offset and energy attenuation are both greater than the preset abnormal threshold is selected as the abnormal area. The sensing unit obtains the volume of the external object based on the magnitude of the impedance shift in the abnormal region in the situation space, combined with the coverage area of ​​the energy attenuation shadow feature in the situation space, and according to the preset binary feature calibration relationship; it tracks the spatial coordinate changes of the impedance and energy attenuation features corresponding to the abnormal region in the time-series continuous fluctuation feature, and obtains the speed of the external object's movement by combining the interval of the real-time electromagnetic signal acquisition time; and it locates the position of the external object within the preset warning range based on the situation space coordinates corresponding to the abnormal region.

6. A computer host protection system according to claim 3 or 5, characterized in that: The sensing unit sends warning information to the protection unit. The protection unit randomly generates an interference trajectory based on the position and speed of the external object in the warning information. The risk value corresponding to any point on the interference trajectory is less than the preset collision risk value. The excitation current is adjusted according to the relationship between the amplitude of the excitation current, the working frequency and the spatial distance, and the excitation current timing trigger interval and the spatial distance of the external object. The protection unit combines the dynamic changes of the excitation current with the electromagnetic fingerprint reference to extract the electromagnetic fingerprint reference corresponding to the interference trajectory and construct an electromagnetic interference fingerprint. The protection unit compares the electromagnetic interference fingerprint with the multi-dimensional feature differences in the situation space and obtains the motion trajectory of the external object based on the feature differences.

7. A computer host protection system according to claim 5, characterized in that: The sensing unit collects the amplitude of the signal loop impedance characteristics at the chassis connection gap and the corresponding time-series change period; the sensing unit extracts the vibration amplitude between the chassis side walls at the chassis gap from the impedance characteristic amplitude, and extracts the vibration period of the chassis side walls at the chassis gap from the time-series change period of the impedance characteristic amplitude. The protection unit obtains the inherent correlation between the impedance value at the chassis gap and the gap width from the electromagnetic fingerprint reference, establishes an impedance gap mapping relationship, substitutes the real-time collected impedance value at the chassis gap into the impedance gap mapping relationship, and obtains the real-time gap width at the chassis connection gap. The protection unit adjusts the current amplitude and phase of the signal circuit corresponding to the chassis gap based on the real-time expansion or contraction trend of the gap width. The protection unit adjusts the current phase, duration of current phase action, and switching sequence of the magnetic induction device on the corresponding signal circuit according to the vibration cycle. When the width of the chassis gap shows an expanding trend, it controls the adjacent magnetic induction devices to form an electromagnetic attraction effect. When the width of the chassis gap shows a contracting trend, it controls the adjacent magnetic induction devices to form an electromagnetic repulsion effect.

8. A computer host protection system according to claim 7, characterized in that: Based on the impedance gap mapping relationship, the sensing unit obtains the real-time gap width corresponding to each connection gap of the chassis, establishes a spatial coordinate system with the geometric center of the chassis as the origin, obtains the geometric position of each connection gap in the spatial coordinate system as the gap position according to the chassis geometric dimensions sent by the maintenance terminal, and stores the gap position in association with the real-time gap width. When the protection unit receives the installation and locking command sent by the operation and maintenance terminal, it acquires and stores the real-time gap width, gap position and movement reference data, and uses the origin coordinates of the spatial coordinate system as the initial centroid coordinates. The protection unit monitors the changes in the width of each gap in real time, adjusts the initial center of gravity coordinates based on the gap position to obtain the moving center of gravity coordinates, and obtains the moving orientation and speed of the moving center of gravity coordinates in the spatial coordinate system based on the acquisition time of the real-time gap width. The distance between the moving center of gravity coordinates and the initial center of gravity coordinates is obtained as the offset distance, and the vibration verification cycle is extracted from the moving orientation and speed. When the difference between the vibration verification period and the vibration period is less than the preset difference threshold and the offset distance is less than the preset stability threshold, the current amplitude and current phase of the signal circuit are adjusted according to the real-time gap width change trend. The protection unit is also used to send a preset movement signal as protection information to the operation and maintenance terminal when the offset distance is not less than the preset offset stability threshold, or when the difference between adjacent vibration verification cycles is greater than the preset vibration stability threshold.

9. A computer host protection system according to claim 8, characterized in that: The sensing unit is also used to collect the real-time temperature inside the chassis, and to perform amplitude compensation on the real-time electromagnetic signal according to the preset temperature sensitivity correction curve. The amplitude-compensated real-time electromagnetic signal is then used for static distribution extraction and acquisition of real-time evolution slope values. When the difference between the vibration verification period and the vibration period is less than the preset difference threshold, and the offset distance is less than the preset stability threshold, the protection unit will combine the combination of vibration period and vibration amplitude with the corresponding real-time temperature to establish a temperature-vibration correlation relationship. Before adjusting the current amplitude and phase of the signal circuit according to the real-time gap width change trend, the combination of the current vibration period and vibration amplitude is verified through the temperature vibration correlation. Based on the verification results, the preset moving signal is sent to the operation and maintenance terminal as protection information.

10. A computer host protective chassis, characterized in that, A computer host protection system according to any one of claims 1-9 was used.

Images