Method and device for tamper detection using anti-tamper radio (ATR)

Abstract

The invention relates to a method and a device for detecting physical tampering on closed systems using anti-tamper radio (ATR). The objective of the present invention is to improve the performance of known ATR systems so that they operate more securely against tampering, are less prone to errors, and are more resource-efficient. To this end, the invention proposes that a reconfigurable intelligent surface (RIS), consisting of at least one controllable reflective element, be provided, the reflective properties of which can be modified by means of an ATR module of the ATR system.

Description

[0001] The present invention relates to a method and a device for detecting physical manipulations on closed systems using anti-tamper radio (ATR). Background of the invention

[0002] Modern information systems store and process highly sensitive data, including cryptographic keys and proprietary intellectual property. While existing security mechanisms primarily target remote attacks, devices are often insufficiently protected against physical tampering. Anti-tamper technologies aim to close this gap by detecting and responding to physical attacks.

[0003] Numerous real-world attack scenarios demonstrate the need for robust anti-tamper measures. For example, the Snowden documents (https: / / www.spiegel.de / international / world / cataloqg-reveals-nsa-has-back-doorsfor-numerous-devices-a-940994.html) revealed that state actors were deliberately compromising hardware components in the supply chain to enable espionage operations. These so-called supply chain attacks involve the undetected modification or replacement of hardware components before they reach the end customer. A well-known example is the so-called "Big Hack" (https: / / www.bloomberq.com / news / features / 2018-10-04 / the-biq-hack-how-china-used-a-tiny-chip-to-infiltrate-america-s-topcompanies), in which tiny chips were allegedly placed in servers of US technology companies to give attackers access to networks and data. Such attacks are difficult to detect because they often leave no noticeable software traces.

[0004] In addition to state actors, cybercriminals and competitors also use physical attacks to steal sensitive information or manipulate critical systems. Methods such as side-channel attacks, which read electromagnetic emissions or power consumption anomalies, or invasive attacks using fine needles to directly scan data lines, are among the known threats. The manipulation of firmware or hardware by injected Trojans also remains a serious threat to security-critical systems.

[0005] The publication “Anti-Tamper Radio: System-Level Tamper Detection for Computing Systems” (P. Staat et al., 2022 IEEE Symposium on Security and Privacy (SP), DOI: 10.1109 / SP46214.2022.00067) describes a method and device mentioned above using an ATR module. An anti-tamper radio (ATR) is a security system that uses radio waves to detect physical tampering with electronic devices. It is based on the continuous monitoring of radio channel characteristics between antennas located within or near the protected system. Initially, a reference measurement is taken to serve as a baseline. During operation, new measurements are regularly compared with this reference to detect even the smallest changes in the environment.Tampering with the device, such as opening a casing or inserting a probe, alters the reflections and attenuation of radio waves, causing the ATR system to detect deviations. If an attack is detected, the system can automatically initiate countermeasures, such as triggering alarms, encrypting data, or locking the device.

[0006] Such a method is also described in DE 10 2017 114 010 A1. WO 2012 / 122994 A1 describes a similar system at the chip level.

[0007] The method behind the Anti-Tamper Radio (ATR) is based on analyzing radio channel characteristics to detect tampering with a protected system. First, a reference measurement is performed, during which the system records the normal radio wave characteristics under undisturbed conditions. These measurements serve as the basis for subsequent comparisons. During operation, the ATR continuously or at regular intervals measures the radio channel characteristics between permanently installed antennas and compares the current values ​​with the stored reference. Parameters such as the signal's amplitude, phase, and frequency response are analyzed.

[0008] If tampering occurs, for example by opening a housing or inserting a metallic object, the reflections, attenuation, or travel times of the radio waves change. These changes are quantified using a distance metric that determines deviations between the current measurement and the reference. If the deviation exceeds a defined threshold, tampering is detected. Task:

[0009] Despite the promising properties of ATR technology, key challenges remain: Although ATR systems are designed to detect physical tampering, there is a risk that attackers could replicate the reflection pattern of the closed enclosure. By using specially adapted reflective materials or active signal generators, an electromagnetic environment could be created that the ATR system cannot identify as tampering. This would allow attackers to make undetected changes to the system or insert eavesdropping mechanisms. Additionally, targeted spoofing or jamming attacks can modify the ATR signal to mask tampering. Therefore, dynamic adjustment of the reflection characteristics is necessary to prevent such attacks.

[0010] Another problem with conventional ATR systems is their high and inconsistent sensitivity to natural environmental changes. Temperature, humidity, or pressure fluctuations within a closed enclosure can alter the electromagnetic wave propagation patterns. Such natural fluctuations can be misinterpreted as manipulation, leading to an increased number of false alarms. In addition to these external environmental influences, internal sources of interference can also affect the measurement results. Electrical and magnetic interference from components such as fans, hard drives, or power supplies can generate unwanted disturbances in the received ATR signals. Furthermore, dynamic changes in the internal structure, for example, due to moving parts or altered temperature distributions, can also result in deviations in the electromagnetic reflection properties.Furthermore, normal operating processes cause periodic changes in the radio environment, for example through high-frequency switching or fluctuating load conditions within the system. Since a secure system must minimize false alarms, it is necessary to differentiate between genuine tampering attempts, internal interference sources, and natural fluctuations.

[0011] Existing ATR systems often require very large bandwidths, frequently in the range of several gigahertz (up to 7 GHz), to reliably detect tampering. However, this presents significant challenges: regulatory restrictions limit the permissible use of electromagnetic signals, complicating implementation. Furthermore, high bandwidths increase the likelihood of interference with other wireless communication systems, particularly in high-density environments such as data centers or IoT environments. Another important aspect is the high energy consumption associated with using large frequency ranges, which is especially problematic in battery-powered or resource-limited systems. Therefore, there is a need to optimize frequency usage to ensure high detection accuracy with the smallest possible bandwidth.

[0012] The object of the present invention is to improve the performance of known ATR systems so that they are more tamper-proof, less prone to errors and more resource-efficient.

[0013] To solve this problem, the invention proposes a method for detecting physical manipulations on a closed system, in particular on electronic devices with security requirements, in which - an anti-tamper radio (ATR) module emits at least one signal within a housing of the system in the form of electromagnetic waves by means of a transmitting unit and receives at least one reflected signal for analysis by means of a receiving unit, - by means of a Reconfigurable Intelligent Surface, which is arranged inside the housing or on the inner wall of the housing of the system, the wave propagation within the housing is dynamically influenced by changing the variable reflection parameters of the RIS by means of a signal processing unit, randomly, periodically, according to a predefined scheme and / or event-controlled, - the ATR module stores at least one reference signature based on the measurement of an initial reference signal, which describes the electromagnetic response of the closed system in an unmanipulated state, - a signature is generated based on the measured signal at each subsequent measurement and compared with the stored reference signature, and - manipulation is detected if the signature received by the ATR module deviates from the reference signature by a predefined threshold.

[0014] Furthermore, the invention proposes a device for detecting physical manipulations on a closed system, in particular on safety-critical electronic devices, comprising - a housing and an anti-tamper radio module arranged within the housing, comprising a transmitter unit, a receiver unit, and a signal processing unit, designed to transmit electromagnetic signals and to detect and evaluate the signals reflected within the housing, and - a Reconfigurable Intelligent Surface consisting of at least one controllable reflective element, arranged in the housing or on the inner wall of the housing and communicatively connected to the signal processing unit, - wherein the signal processing unit (C) is configured to change the reflection properties of the RIS randomly, periodically, according to a predefined scheme and / or event-driven in order to make signal manipulation by an attacker (A1) more difficult.

[0015] RIS technology is used to dynamically control the propagation of electromagnetic waves. In the prior art, RIS technology is currently used in wireless communication to improve signal strength and quality. One example is patent application US20220337240A1, which describes an electromagnetic-circuit co-design method for highly reconfigurable, multifunctional, and high-speed programmable metasurfaces with integrated chip tiling.

[0016] While known RIS applications mainly aim to optimize communication channels, the present invention uses RIS to improve performance through integration into ATR systems.

[0017] The RIS improves tamper detection by dynamically altering the reflection properties of electromagnetic waves within the enclosure. This keeps the reflection pattern unpredictable and prevents attackers from replicating it. Spoofing and jamming attacks are made more difficult because the RIS can actively respond to external tampering attempts and adjust the reflection behavior to detect deception.

[0018] To reduce false alarms, the RIS reflection parameters are adaptively adjusted to changes in the environment. Internal sources of interference, such as fans or power supplies, as well as environmental influences like temperature and humidity fluctuations, are compensated for. This ensures a more stable electromagnetic environment, so the ATR system only reacts to genuine tampering attempts and not to natural fluctuations or regular operating processes. The false positive rate is significantly reduced.

[0019] Frequency band utilization is optimized by RIS by selectively generating reflection patterns that enable unambiguous tamper detection even with narrower bandwidths. This allows ATR to operate with less spectrum, circumvent regulatory restrictions, and avoid interference with other radio systems. Simultaneously, energy consumption is reduced because broadband signal analysis is no longer required.

[0020] One application example would be a secure data center with servers where casing openings and tampering attempts need to be detected. Conventional methods such as mechanical sensors can be circumvented through targeted interventions. However, by using ATR with RIS, any tampering, even a small opening of the casing, is made detectable through changes in wave propagation.

[0021] In the sense of the present invention, a housing encloses at least partially protectable hardware components and thus defines an environment.

[0022] The transmitting unit and the receiving unit can also be formed by a single module, for example a transceiver.

[0023] The signal processing unit can be designed, for example, as a microcontroller or FGPA, and is set up to control the RIS in order to change its reflection properties.

[0024] Furthermore, it is proposed that the ATR module performs measurements periodically or event-driven to continuously detect tampering attempts and ensure that changes in the electromagnetic signature do not go unnoticed. A static tampering detection system would be vulnerable to attacks, as an attacker could operate undisturbed after successfully bypassing the detection. Therefore, it is advantageous for the ATR system to perform measurements either regularly (e.g., every few minutes) or in response to external events (e.g., suspected tampering). For example, a high-security safe, a server, or critical OT equipment can be continuously monitored using this technique. If an attacker attempts to open the enclosure undetected, this will be detected by a periodic ATR measurement.If an attacker has opened the casing unnoticed during a power outage and made changes to the system, this will be detected by an ATR measurement as soon as power is restored.

[0025] Furthermore, it is proposed that the RIS be given randomly varying reflection parameters, making the system's electromagnetic response unpredictable for a potential attacker. The system's electromagnetic response is not constant, thus preventing unauthorized replication of the environment by an attacker. A significant disadvantage of conventional ATR systems is that an attacker could theoretically create an exact copy of the electromagnetic environment. Using a dynamically changing RIS prevents this, as the reflection conditions are constantly changing. An example is the monitoring of a military weapons depot: random changes in the RIS's reflection characteristics prevent an attacker from creating a static electromagnetic environment to conceal tampering.

[0026] Furthermore, it is proposed that the RIS modifies the configuration of the reflection elements in a predefined sequence or using a random algorithm to ensure that no static electromagnetic signature is generated that could cause strong interference or environmental influences. These changes can occur not only randomly but also according to specific patterns. This can be helpful in accounting for environmental changes or legitimate sources of interference, such as fans or rotating hard drives, or in performing targeted interference signal analysis. An example would be server security: by regularly changing the reflection properties, legitimate sources of interference can be detected and false positive attack detections reduced.

[0027] Furthermore, it is proposed that the received signals be analyzed using digital signal processing methods, preferably Euclidean distance metrics, to detect even the slightest changes in the reflection signature and thereby reliably detect particularly subtle manipulations such as the insertion of microprobes or the unauthorized insertion of components. Manipulations of highly sensitive devices, for example, the insertion of a listening device into a cryptography module, can be very difficult to detect. High-precision digital signal processing can detect even the insertion of a thin probe.

[0028] Furthermore, it is proposed that artificial intelligence (AI)-supported pattern analysis be used to distinguish manipulation-induced signature deviations from natural signature deviations caused by system- or environmental disturbances. This would allow legitimate variations, such as those caused by environmental influences like temperature fluctuations or changes in humidity, to be differentiated from actual manipulation attempts. A problem with sensor-based manipulation detection is the high error rate due to environmental influences. An AI can learn from training data to distinguish legitimate changes from genuine threats.

[0029] Furthermore, it is proposed that in the event of tampering, the system initiates an alarm or protective measure, preferably by locking security-critical data, sending a notification to a monitoring center, or triggering a physical security protocol. Tamper detection alone may not be sufficient, and a direct security response may be necessary. For example, a smart meter could be automatically locked by tampering, or a high-security server could encrypt all data in the event of an alarm.

[0030] Furthermore, it is proposed that the bandwidth used for measurement be limited to less than 100 MHz, preferably 20 MHz, through dynamic control of the RIS. Reducing the required bandwidth from 7 GHz to less than 100 MHz or 20 MHz enables more efficient use of the radio spectrum. Reducing the bandwidth to less than 100 MHz through dynamic RIS control offers several crucial advantages. Because the system operates with a lower bandwidth, interference with other radio systems is minimized, especially in environments with high frequency utilization such as data centers or IoT networks. At the same time, the reduced bandwidth usage facilitates regulatory compliance, as many frequency ranges are subject to legal restrictions. This also enables more flexible integration into existing systems.Another advantage lies in the reduced energy consumption, as less spectrum is used for signal processing. This is particularly important for battery-powered or energy-efficient applications. Despite the reduced bandwidth, detection accuracy is maintained because the RIS selectively modifies and adapts the reflection patterns within the housing to reliably detect tampering attempts. This ensures robust and precise detection even with reduced available bandwidth. Additionally, the dynamic control of the RIS increases the efficiency of the ATR system, as only the actually required frequency resources are used, without unnecessarily burdening the radio spectrum. Overall, the bandwidth reduction provided by the RIS enables energy-efficient, low-interference, and regulatory-compliant tamper detection without compromising the system's security or reliability.

[0031] According to the invention, the signal processing unit modifies the reflection properties of the RIS randomly, periodically, according to a predefined scheme, and / or event-driven to make signal manipulation by an attacker more difficult. This ensures continuous variability of the electromagnetic field and thus minimizes the risk of successful manipulation by mimicking the environment. Manipulation attempts can extend over long periods, with attackers gradually analyzing and mimicking the electromagnetic environment of a system to deceive the ATR. However, due to the continuous adjustment of the RIS's reflection properties, the electromagnetic response remains unpredictable.

[0032] For example, an insider attacker might target server racks in a data center to measure electromagnetic reflections over an extended period and generate a spoofing signal. However, if the RIS (Reflection Array of Signals) regularly changes its reflection characteristics, such an attack technique will be ineffective because the patterns are unpredictable.

[0033] Furthermore, it is proposed that the signal processing unit be equipped with a machine learning algorithm that can distinguish between natural environmental influences and actual tampering attempts in order to reduce false alarms and ensure more reliable security monitoring. One of the challenges with sensor-based security systems is the differentiation between natural changes and actual tampering attempts. For example, temperature changes, humidity, or moving internal components such as fans or cables can cause a change in the electromagnetic field without any tampering occurring. The invention therefore uses machine learning (ML) to analyze patterns in the electromagnetic reflections and to recognize whether a change is due to normal environmental conditions or to unauthorized tampering. This significantly improves the false-positive rate.

[0034] A practical example would be the monitoring of an OT system in the water supply, which operates over years under varying climatic conditions. The ML component learns to distinguish legitimate changes, such as seasonal temperature fluctuations, from actual interventions.

[0035] Furthermore, it is proposed that the RIS be either integrated directly into the inner wall of the enclosed housing or mounted as a separate component within the system to ensure particularly high tamper resistance. The positioning of the RIS plays a crucial role in the effectiveness of tamper detection. Integrating the RIS into the inner wall of the housing creates a particularly tamper-resistant system, as attackers have no way to deactivate or bypass it without damaging the housing itself.

[0036] This is particularly relevant for highly secure industrial systems or banks, where attackers could attempt to deceive the ATR using micro-tools. An integrated RIS would electromagnetically detect even small openings in the casing or changes inside, recognizing them as tampering.

[0037] Furthermore, it is proposed that the device include a data processing unit that stores manipulation events in a tamper-proof database or blockchain to enable subsequent forensic analysis and auditable documentation of the security history. An effective security concept requires traceable documentation of manipulation events. By storing ATR measurements and changes in a tamper-proof database or blockchain, previous attack attempts can be analyzed and security measures adjusted.

[0038] One application example is a safety-critical data center where any changes in electromagnetic reflections are documented over months. This makes it possible to track whether there are patterns of suspicious changes that indicate targeted manipulation attempts over extended periods.

[0039] Furthermore, it is proposed that the ATR system operate in a frequency range between 2 GHz and 9 GHz to ensure an optimal balance between sensitivity to tampering attempts and resistance to external interference. The selected frequency range between 2 GHz and 9 GHz was specifically chosen to guarantee an optimal balance between penetration depth into the enclosure and sensitivity to tampering. Lower frequencies below 2 GHz could lead to significant wave propagation and reduced tamper detection, while higher frequencies above 9 GHz are susceptible to environmental influences and reflection losses.

[0040] Furthermore, it is proposed that the manipulation detection enables real-time alerting to an external monitoring center, such as a Security Operating Center (SOC), a control room, or a central IT security infrastructure, so that countermeasures can be initiated immediately in the event of detected manipulation. A significant advantage of the invention lies in the real-time monitoring and alerting. In security-critical areas such as banks, high-security servers, or government institutions, it is not sufficient to merely log manipulations—they must be detected and reported immediately.

[0041] For example, a control unit in a car could immediately trigger an alarm if the reflection behavior changes suddenly or after an offline period, indicating an attempted tampering. This would not only activate the alarm but also inform the monitoring systems and initiate possible countermeasures (e.g., blocking cryptographic remote access, activating verification measures).

[0042] Another example would be securing a blockchain server: If manipulation is detected, the system could automatically lock critical keys or pause transactions to prevent greater damage.

[0043] The device and method according to the invention can be integrated into existing security architectures without requiring any physical modifications to the device. Examples of implementation

[0044] Exemplary embodiments of the invention are explained in more detail below with reference to drawings and graphs. These show: Fig. 1: schematically a device according to the invention in an embodiment for manipulation detection using ATR and RIS on the example of a server; Fig. 2: the device made of Fig. 1 with wave propagation; Fig. 3: the device made of Fig. 1 in case of attack; Fig. 4: Block diagram for the sequence of the method according to the invention; Fig. 5: schematically the detection rate of a method according to the invention over the bandwidth; Fig. 6: Intra-distance measurements over time under different conditions.

[0045] In Fig. Figure 1 shows an embodiment of a device according to the invention. The device comprises an anti-tamper radio module (ATR) and a reconfigurable intelligent surface (RIS) arranged within a protected enclosure (G). The enclosure (G) is depicted as a rectangular box, symbolizing a secure system such as a server or a hardware security module. It serves as physical protection for sensitive electronic components and prevents unauthorized access. It includes, by way of example, several fans (L), a mainboard (E1), a graphics card (E2), a network card (E3), and a hard drive (M), which together represent the information processing system (D) to be protected.

[0046] The ATR module comprises a signal processing unit C, a transmitter T, and a receiver R. The transmitter T is connected to a small antenna T1, which emits electromagnetic waves (radio waves), while the receiver R uses an antenna R1 to detect the reflected signals.

[0047] To precisely control the reflection of radio waves, the RIS (Radio Frequency Identification) is mounted on the inner walls of the enclosure. The RIS consists of several small rectangular elements that modify the reflection of radio waves by dynamically adjusting their surface properties. The RIS can be controlled via a communication link S using the signal processing unit. This configuration ensures that the electromagnetic environment is not static, thus preventing attackers from replicating a repeatable signature.

[0048] In Fig. 2 are compared to the Fig. 1 The radio waves W are visualized by dashed arrows, which represent the propagation of radio waves within the entire housing G.

[0049] In Fig. 3 is compared to the Fig. Figure 2 depicts a manipulation attempt by an attacker A1. The attack occurs, for example, through the penetration of a tool A3 or the opening of the housing G. The tool A3 could be, for example, a robotic arm with a fine needle, a screwdriver, or a thin probe. Using tool A3, the attacker attempts to gain access to the interior of the housing G and, in particular, to the sensitive hardware components E1-E3. This leads to a change in the wave propagation of the radio waves W, which is highlighted in the diagram by differing reflection paths and dotted arrows. This deviation signals manipulation of the protected system. The attacker attempts to compensate for this change by introducing their own signals A2. However, due to the dynamic change in the reflection parameters of the RIS, the attacker's attempt fails.

[0050] Fig. Figure 3 also visualizes a signal analysis and alarm mechanism implemented by the ATR. The signal processing unit C in the ATR module detects deviations in the signature of the reflected radio waves W, analyzes them for tampering indicators, and then alerts internal and / or external systems, such as a Security Operations Center (SOC). The connection between the ATR module and the respective system is represented by data lines B1 and B2 to illustrate the data flow from the ATR module to the security system.

[0051] Fig. Figure 4 illustrates the process of the inventive method for tamper detection using ATR and RIS. The process begins with the initialization of the system, in which all components are returned to their initial state. Subsequently, the reference measurement is performed, in which the ATR module emits electromagnetic signals that are reflected by the RIS.

[0052] The receiving unit captures these reflections and stores the resulting signature as a reference signature, which describes the unaltered, tamper-proof state of the protected system. Multiple reference signatures can also be created, each associated with different reflection parameter settings of the RIS.

[0053] After the reference measurement, the system switches to operating mode, in which continuous monitoring takes place. The ATR module transmits radio signals at regular intervals or in response to external events. Simultaneously, the RIS continuously changes its reflection pattern to ensure that no static electromagnetic field is generated that could be replicated by an attacker.

[0054] In the next stage, the current measurement is compared with the reference data. The signal processing unit analyzes the reflected signals and compares them with the stored reference signature. To prevent false alarms caused by natural environmental changes such as temperature fluctuations or mechanical vibrations, a machine learning algorithm is used to filter out environmental factors and identify only safety-critical deviations. This process continues until tampering or alteration is detected, or the system is shut down. Upon restarting, the system begins with this third step.

[0055] If tampering is detected, the system immediately activates the alarm system. Otherwise, monitoring continues without interruption. If a deviation is detected, various security measures can be triggered, including the encryption or blocking of sensitive data, notification of a monitoring center, or the activation of physical safeguards such as shutting down the system or activating mechanical locks. The process then ends.

[0056] Figure 5 shows an example of the bandwidth-dependent detection rate of an ATR module under conditions where the RIS is both enabled and disabled. The results show that the RIS significantly reduces the required bandwidth, such that a 20 MHz bandwidth in combination with the RIS is equivalent to a 7 GHz bandwidth without the RIS. Thus, the use of the RIS according to the invention results in a considerably lower bandwidth requirement.

[0057] Figure 6 shows an example of the intra-distance measurements on the device made of Fig. 1. Under various conditions, the system was tested: firstly, with the RIS (Remote Sensor Operation) switched off, both with the fan switched on and off; and secondly, with the RIS switched on while the fan was running. The results show that optimizing the RIS significantly reduces the influence of the fan, bringing the system response closer to the condition under which the fan is switched off. The intra-distance refers to the signals measured by the ATR (Automatic Temperature Regulator) module, which represent the environmental response in the absence of manipulation. Thus, false alarms due to legitimate changes—in this case, fan operation—are significantly reduced because the deviations remain minor and do not exceed a predefined threshold. Reference symbol list: G Housing / Protected System ATR Anti-Tamper Radio Module W radio wave propagation C Signal processing unit T transmitting unit of the ATR T1 antenna of the transmitter unit R receiver unit of the ATR R1 antenna of the receiving unit RIS Reconfigurable Intelligent Surface (RIS) S Communication link from the ATR module to the RIS L fan M hard drive E1 Mainboard E2 Graphics Card (GPU) E3 network card A1 Attacker A2 Jamming signal to compensate for the effects of attack A3 A3 Attack B1 Data connection B2 Data connection SOC Security Operation Center

Claims

[1] Method for detecting physical tampering with a closed system, in particular with electronic devices with security requirements, in which - an anti-tamper radio (ATR) module emits at least one signal within a housing (G) of the system in the form of electromagnetic waves by means of a transmitting unit (T) and receives at least one reflected signal for analysis by means of a receiving unit (R), characterized by ,that - by means of a Reconfigurable Intelligent Surface (RIS) which is arranged inside the housing (G) or on the inner wall of the housing (G) of the system, the wave propagation within the housing (G) is dynamically influenced by changing the variable reflection parameters of the RIS by means of a signal processing unit (C), randomly, periodically, according to a predefined scheme and / or event-driven, - the ATR module (ATR) stores at least one reference signature based on the measurement of an initial reference signal, which describes the electromagnetic response of the closed system in an unmanipulated state, - a signature is generated based on the measured signal at each subsequent measurement and compared with the stored reference signature, and - manipulation is detected if the signature received by the ATR module deviates from the reference signature by a predefined threshold. [2] Method according to claim 1, characterized by , that the ATR module (ATR) performs periodic or event-driven measurements to continuously detect tampering attempts. [3] Method according to any one of the preceding claims, characterized by , that the RIS changes the configuration of the reflection elements in a predefined sequence or by a random algorithm. [4] Method according to any one of the preceding claims, characterized by that the received signals are analyzed using methods of digital signal processing, preferably the Euclidean distance metric, in order to detect even subtle manipulations. [5] Method according to any one of the preceding claims, characterized by , that in addition, an artificial intelligence (AI)-supported pattern analysis is used to distinguish manipulation-induced signature deviations from natural signature deviations caused by system- or environment-related disturbances. [6] Method according to any one of the preceding claims, characterized by , that in the event of manipulation the system initiates an alarm or protective measure, preferably locking sensitive data, activating a physical security mechanism or notifying a monitoring center (SOC). [7] Method according to any of the aforementioned claims, characterized bythat the bandwidth used for the measurement is limited to less than 1 GHz, preferably to 100 MHz, and particularly preferably to 20 MHz, by the dynamic control of the RIS. [8] Device for detecting physical tampering with a closed system, in particular safety-critical electronic equipment, comprising - a housing (G) and an anti-tamper radio (ATR) module arranged in the housing (G) comprising a transmitter unit (T), a receiver unit (R) and a signal processing unit (C), configured to emit electromagnetic signals and to detect and evaluate the signals reflected within the housing, characterized by , that - a Reconfigurable Intelligent Surface (RIS) consisting of at least one controllable reflective element, arranged in the housing (G) or on the inner wall of the housing (G) and communicatively connected to the signal processing unit (C), - wherein the signal processing unit (C) is configured to change the reflection properties of the RIS randomly, periodically, according to a predefined scheme and / or event-driven in order to make signal manipulation by an attacker (A1) more difficult. [9] Device according to claim 8, characterized by , that the signal processing unit (C) is equipped with a machine learning algorithm that can distinguish between legitimate environmental influences and actual manipulation attempts. [10] Device according to one of claims 8-9, characterized by , that the RIS is integrated into the inner wall of the housing (G) or is installed as a separate component within the housing (G). [11] Device according to one of claims 8-10, characterized by, that a data processing unit is connected to the signal processing unit (C), which logs manipulation events and preferably stores them in a blockchain or a tamper-proof database. [12] Device according to one of claims 8-11, characterized by , that the ATR module is set up to operate in a frequency range between 2 GHz and 9 GHz to ensure high sensitivity to tampering attempts. [13] Device according to one of claims 8-12, characterized by , that the signal processing unit is set up to enable real-time alerting to an external monitoring center (SOC) or a central IT security infrastructure in the event of tampering detection.

Images