A method and system for covert communication in a cellular internet of things

Abstract

The application provides a kind of hidden communication method and system of cellular internet of things, it is related to wireless communication security technical field.The method is by CPU scheduling internet of things device as Alice and Bob, obtains the first link CSI between Alice and each AP and the second link CSI between each AP and Bob, and according to the first link CSI and preset link gain transmission threshold, candidate AP set is filtered;According to the statistical CSI of Alice to Eve eavesdropping link and mode switching threshold, determine the probability that Eve switches between HD and FD, and determine the concealment constraint and Bob reliability constraint;Joint optimization is carried out to preset link gain transmission threshold, relay AP quantity and predetermined energy threshold, and the parameter combination for optimizing Bob concealment rate is determined;According to parameter combination, determine relay AP set, adopt relay auxiliary CIPC strategy to determine Alice transmitting power and relay AP forwarding power, and perform hidden communication through relay AP, so as to take into account the communication behavior concealment, the location concealment of both sides of communication and Bob receiving reliability.

Description

Technical Field

[0001] This invention relates to the field of wireless communication security technology, specifically to a non-cellular Internet of Things (IoT) covert communication method and system. Background Technology

[0002] Cellular-free IoT provides communication services to IoT devices by coordinating multiple distributed access points (APs), mitigating the impact of cell boundaries on communication quality at the edge devices in traditional cellular networks, and optimizing resource allocation and scheduling based on the actual needs of connected devices. However, while the distributed architecture of cellular-free IoT improves coverage and service efficiency, it also exposes wireless transmission to security risks inherent in open channels. With increasing capabilities for eavesdropping detection, listening, and interference, the exchange of large amounts of sensitive information between IoT devices is more likely to lead to privacy breaches.

[0003] Existing data security transmission technologies typically include encryption and physical layer security. Encryption primarily protects transmitted content by increasing decoding complexity, while physical layer security mainly utilizes the characteristics of wireless channels to reduce the likelihood of eavesdroppers decoding communication information. Covert communication technologies further focus on whether communication activities are detected by eavesdroppers, and can reduce the probability of legitimate communication being detected and identified to a certain extent. Currently, there are related solutions for IoT covert communication, such as a scheme that achieves covert modulation through spectrum detection and spectrum energy modulation to reduce the energy consumption and cost of IoT covert communication; a covert communication scheme proposed to address the problem of easy eavesdropping and identification of wireless communication in the power IoT; and a scheme that combines reconfigurable smart reflectors, multi-antenna input / output, and deep reinforcement learning to construct a covert communication model for railway IoT. The above solutions improve the covertness or reliability of IoT communication from different perspectives, but overall they still mainly focus on protecting communication behavior or content.

[0004] In non-cellular IoT scenarios, access points (APs) are typically deployed publicly, allowing eavesdroppers to obtain AP location information and the approximate location of IoT devices. If the eavesdropper further deduces the location of the transmitter or receiver, they may reduce the reception quality of legitimate receivers through link observation or artificial noise interference, or even cause legitimate communication to be interrupted. Existing covert communication solutions do not adequately consider the privacy protection of the location of both communicating parties, and lack sufficient adaptation to intelligent eavesdropping scenarios where eavesdroppers adaptively switch between half-duplex eavesdropping mode and full-duplex interference eavesdropping mode based on the state of the eavesdropping link. Therefore, in non-cellular IoT networks with intelligent eavesdroppers like Eve, a covert communication solution that can balance the concealment of communication behavior, the concealment of the location of both communicating parties, and the reliability of legitimate reception is still needed. Summary of the Invention

[0005] The purpose of this invention is to solve the technical problems mentioned above and propose a non-cellular IoT covert communication method, applicable to non-cellular IoT networks where a smart eavesdropper named Eve exists, comprising: The central processing unit (CPU) schedules a pair of IoT devices as Alice, the transmitter, and Bob, the receiver, and notifies multiple access points (APs) to enter the channel estimation state. Obtain the first link channel status information (CSI) between Alice and each AP, and the second link CSI between each AP and Bob; Based on the first link CSI and the preset link gain transmission threshold, candidate APs that meet the transmission conditions are selected from multiple APs to form a candidate AP set, and the probability of Alice transmitting covert information is determined. Based on the CSI statistics and mode switching threshold of the Alice-Eve eavesdropping link, the probability of Eve switching between half-duplex eavesdropping mode (HD) and full-duplex jamming eavesdropping mode (FD) is determined. The concealment constraint is determined based on the average minimum error detection probability of Eve in HD and FD. The reliability constraint of Bob is determined based on the average interference power generated when Eve is in FD. The preset link gain transmission threshold, the number of relay APs, and the predetermined energy threshold are jointly optimized, and the parameter combination used to optimize the Bob concealment rate is determined from the preset link gain transmission threshold, the number of relay APs, and the predetermined energy threshold values ​​that satisfy the concealment constraint and the Bob reliability constraint. The relay AP set is determined from the candidate AP set based on the parameter combination, and the relay auxiliary channel inversion power control (CIPC) strategy is used to determine Alice's transmit power and the relay AP's forwarding power based on the first link CSI, the second link CSI and the predetermined energy threshold, so that the received signals from Alice to the relay AP and from the relay AP to Bob both reach the predetermined energy threshold. Based on Alice's transmit power, relay AP forwarding power, and the relay AP set, relay-assisted covert communication is performed between Alice and Bob.

[0006] In the preferred embodiment, based on the first link CSI and a preset link gain transmission threshold, candidate APs that meet the transmission conditions are selected from multiple APs to form a candidate AP set, including: The first link gain between Alice and each AP is compared with the preset link gain transmission threshold. If the first link gain corresponding to a certain AP is greater than the preset link gain transmission threshold, then the corresponding AP is determined as a candidate AP; If the number of candidate APs is greater than 1 and less than the total number of APs, then the candidate APs are combined into a candidate AP set. If the number of candidate APs does not meet the condition of being greater than 1 and less than the total number of APs, then the current transmission time slot will not initiate covert information transmission, or it will enter the subsequent parameter optimization process to adjust the preset link gain transmission threshold.

[0007] In the preferred scheme, a relay-assisted CIPC strategy is used to determine the Alice transmit power and the relay AP forwarding power, including: Based on the first link gain between Alice and any relay AP, the predetermined energy threshold, and the power allocation factor allocated to any relay AP, determine the transmit power of Alice so that the energy of the received signal at any relay AP reaches the predetermined energy threshold. Based on the second link gain between any relay AP and Bob and a predetermined energy threshold, determine the forwarding power of any relay AP so that the energy of the signal Bob receives from any relay AP reaches the predetermined energy threshold. The energy of the signal received at any relay AP is determined by Alice's transmit power, power allocation factor, and corresponding first link gain. The energy of the signal received by Bob from any relay AP is determined by the forwarding power of any relay AP and corresponding second link gain.

[0008] In the preferred embodiment, the probability of Alice transmitting covert information is determined based on the first link CSI and a preset link gain transmission threshold, including: Based on the number of APs in the current transmission time slot whose first link gain is greater than the preset link gain transmission threshold, determine whether the current transmission time slot meets the probabilistic transmission condition; When the number of APs whose first link gain is greater than the preset link gain transmission threshold is at least two and less than the total number of APs, the current transmission time slot is determined to meet the probabilistic transmission condition. When all first links use the same statistical parameters, the probability of Alice initiating covert information transmission is determined based on the total number of APs, the statistical parameters corresponding to the first link gain, and the preset link gain transmission threshold. When different statistical parameters are used for each first link, the probability of each AP meeting the preset link gain transmission threshold is determined. By enumerating the AP combinations that meet the transmission conditions and accumulating the corresponding combination probabilities, the probability of Alice initiating covert information transmission is determined.

[0009] In the preferred scheme, determining the probability of Eve switching between HD and FD includes: When the gain of the Alice-Eve eavesdropping link is greater than the mode switching threshold, it is determined that Eve has switched to HD and is using two antennas to eavesdrop. When the gain of the Alice-Eve eavesdropping link is less than the mode switching threshold, it is determined that Eve has switched to FD and uses one antenna to send artificial noise and the other antenna to eavesdrop. Based on the CSI statistics and mode switching threshold of the Alice-Eve eavesdropping link, determine the probability that Eve is in FD.

[0010] In the preferred scheme, the concealment constraints are determined based on Eve's average minimum false detection probability under HD and FD, including: Based on the current transmission parameters, the probability of Alice transmitting covert information, the probability of Eve being in FD, and the eavesdropping link statistics CSI, determine Eve's average minimum error detection probability under FD and Eve's average minimum error detection probability under HD respectively. The smaller of Eve's average minimum false detection probability under FD and Eve's average minimum false detection probability under HD is used as the concealment evaluation value; Based on the comparison between the concealment evaluation value and the concealment constraint value, it is determined whether the current transmission parameters meet the concealment constraints. The concealment constraint value is determined based on the smaller of the probability that Alice will transmit concealed information and the probability that Alice will not transmit concealed information, and a preset positive number.

[0011] In the preferred scheme, Bob's reliability constraints are determined, including: Based on the predetermined energy threshold, the set of relay APs, the noise power at each relay AP, the noise power at Bob, and the average interference power generated when Eve is in FD, determine whether the predetermined energy threshold simultaneously satisfies the decoding reliability of the relay APs and the receiving reliability of Bob. When the noise power of multiple relay APs is different, each relay AP should be judged separately, or the maximum value among the noise power of each relay AP should be used for conservative judgment.

[0012] In the preferred scheme, a preset link gain transmission threshold, the number of relay APs, and a predetermined energy threshold are jointly optimized. Then, among the values ​​of the preset link gain transmission threshold, the number of relay APs, and the predetermined energy threshold that satisfy both the concealment constraint and the Bob reliability constraint, a parameter combination for optimizing the Bob concealment rate is determined, including: The preset link gain transmission threshold, the number of relay APs, and the predetermined energy threshold are used as parameters to be optimized. Under the condition of a fixed predetermined energy threshold, the values ​​of the preset link gain transmission threshold and the number of relay APs are determined based on the candidate AP screening results and Bob's concealment rate. Based on the predetermined link gain transmission threshold and the number of relay APs, the predetermined energy threshold is updated using gradient descent or alternating optimization methods. After each update of the predetermined energy threshold, a concealment constraint evaluation and a Bob reliability constraint evaluation are performed. The parameter combination used for covert communication is updated based on the values ​​of the preset link gain transmission threshold, the number of relay APs, and the predetermined energy threshold that satisfy the concealment constraint and the Bob reliability constraint, until the Bob concealment rate converges or the preset iteration stopping condition is reached.

[0013] The present invention also provides a non-cellular IoT covert communication system for use in a non-cellular IoT network where a smart eavesdropper Eve exists, including a central processing unit (CPU), multiple access points (APs), and multiple IoT devices. The CPU is used to schedule a pair of IoT devices as Alice and Bob, notify multiple APs to enter the channel estimation state, obtain the first link CSI between Alice and each AP and the second link CSI between each AP and Bob, filter candidate APs based on the first link CSI and the preset link gain transmission threshold to form a candidate AP set, determine the probability of Alice transmitting covert information, determine the probability of Eve switching between half-duplex eavesdropping mode (HD) and full-duplex interference eavesdropping mode (FD) based on the Alice-Eve eavesdropping link statistical CSI and mode switching threshold, determine the covert constraints based on the average minimum error detection probability of Eve in HD and FD, determine the Bob reliability constraints based on the average interference power generated by Eve in FD, jointly optimize the preset link gain transmission threshold, the number of relay APs and the predetermined energy threshold, and determine the parameter combination used to optimize Bob's covert rate from the values ​​of the preset link gain transmission threshold, the number of relay APs and the predetermined energy threshold that satisfy the covert constraints and Bob reliability constraints, determine the relay AP set from the candidate AP set based on the parameter combination, and use the relay-assisted CIPC strategy to determine Alice's transmit power and the relay AP forwarding power. The AP is used to perform link CSI estimation and feedback under CPU control, and when it is identified as a relay AP, it receives covert information sent by Alice, decodes the covert information, and forwards the covert information to Bob. The IoT device is used to send covert information according to the Alice transmit power determined by the CPU when it is configured as Alice by the CPU, and to receive the forwarded signal from the relay AP and restore the covert information when it is configured as Bob by the CPU.

[0014] In the preferred embodiment, the CPU includes a scheduling control module, a CSI management module, a candidate AP screening module, a relay-assisted CIPC power control module, a probability transmission judgment module, an Eve mode evaluation module, a hidden constraint evaluation module, a Bob reliability evaluation module, a parameter optimization module, and a transmission control module. The scheduling control module is used to schedule Alice and Bob and notify the AP to enter the channel estimation state; The CSI management module is used to receive and organize the first link CSI and the second link CSI. The candidate AP filtering module is used to filter candidate APs based on the first link CSI and a preset link gain transmission threshold. The relay-assisted CIPC power control module is used to determine the Alice transmit power and the forwarding power of each relay AP; The probability transmission judgment module is used to determine the probability that Alice will transmit covert information; The Eve mode evaluation module is used to determine the probability that Eve is in FD based on the CSI statistics and mode switching threshold of the Alice-Eve eavesdropping link. The concealment constraint evaluation module is used to evaluate whether the values ​​of the preset link gain transmission threshold, the number of relay APs, and the predetermined energy threshold satisfy the concealment constraints based on Eve's average minimum error detection probability under FD and HD. The Bob reliability evaluation module is used to evaluate whether the values ​​of the preset link gain transmission threshold, the number of relay APs, and the predetermined energy threshold meet the Bob reliability constraints based on the average interference power generated when Eve is in FD. The parameter optimization module is used to jointly optimize the preset link gain transmission threshold, the number of relay APs, and the predetermined energy threshold, and to determine the parameter combination for covert communication from the preset link gain transmission threshold, the number of relay APs, and the predetermined energy threshold values ​​that satisfy the concealment constraint and the Bob reliability constraint. The transmission control module is used to send power control parameters and the set of relay APs to Alice and the corresponding relay APs, so as to control Alice to send covert information to the relay APs and control the relay APs to forward covert information to Bob.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention utilizes CPU scheduling of Alice and Bob to acquire the first and second link CSIs. Based on a preset link gain transmission threshold, it filters a set of candidate access points (APs) and determines the probability of Alice transmitting covert information. This enables a probabilistic covert transmission mechanism based on legitimate link CSIs in non-cellular IoT networks. A relay-assisted CIPC strategy determines Alice's transmit power and relay AP forwarding power, ensuring that the received signals from Alice to the relay AP and from the relay AP to Bob both reach predetermined energy thresholds. This reduces the likelihood of Eve inferring the locations of both communicating parties based on link status and received signal changes, thus balancing the covertness of communication behavior and the location of both parties. Furthermore, this invention determines the probability of Eve switching between half-duplex eavesdropping mode (HD) and full-duplex interference eavesdropping mode (FD) based on the Alice-Eve eavesdropping link statistical CSI and mode switching thresholds. It determines covert constraints based on Eve's average minimum error detection probability in HD and FD, and determines Bob's reliability constraints based on the average interference power generated when Eve is in FD. This allows the transmission parameter evaluation to adapt to the mode switching behavior of the intelligent eavesdropper Eve. Based on this, the preset link gain transmission threshold, the number of relay APs, and the predetermined energy threshold are jointly optimized. The parameter combination used to optimize the Bob covert rate is determined from the parameter values ​​that satisfy the covert constraints and Bob reliability constraints. This can improve the Bob covert rate while ensuring the decoding reliability of relay APs and the receiving reliability of Bob. Attached Figure Description

[0016] The present invention will be further described below with reference to the accompanying drawings and embodiments: Figure 1 A schematic diagram of the overall process of a non-cellular Internet of Things (IoT) covert communication method; Figure 2 This is a schematic diagram of the joint optimization process of probability transmission and relay-assisted CIPC; Figure 3 This is a diagram of a non-cellular Internet of Things (IoT) covert communication system architecture. Detailed Implementation

[0017] Example 1 like Figure 1 As shown, this embodiment provides a covert communication method for non-cellular IoT networks, applied to a non-cellular IoT network with an intelligent eavesdropper named Eve. The network includes a central processing unit (CPU), multiple access points (APs), multiple IoT devices, and Eve. The CPU communicates with each AP via a backhaul link and is responsible for scheduling the IoT devices and APs; each AP is deployed within its coverage area to assist IoT devices in completing covert communication.

[0018] Multiple transmission slots are configured within a single transmission cycle. For any given transmission slot, the CPU schedules a pair of IoT devices from multiple IoT devices, designating one IoT device as the transmitter Alice and the other as the receiver Bob. Other unscheduled IoT devices remain silent during that transmission slot. This configuration limits the legitimate communication objects within a single transmission slot and provides a scheduling basis for subsequent AP filtering, power control, and covert transmission detection.

[0019] Before communication begins, the CPU instructs each AP to perform channel estimation. Each AP estimates the first link CSI between Alice and each AP, and the second link CSI between each AP and Bob, based on the pilot signal, and feeds the estimation results back to the CPU. The first link CSI, used to determine Alice's transmit power or power allocation factor, can be fed back to Alice by the corresponding AP, or sent to Alice by the CPU based on the feedback result. The first link CSI is used to determine whether the links from Alice to each AP meet the covert transmission conditions and to determine Alice's transmit power or power allocation factor; the second link CSI is used to determine the forwarding power when each AP forwards covert information to Bob.

[0020] CPU sets preset link gain transmission threshold And compare the first link gain between Alice and each AP with the preset link gain transmission threshold. A comparison is made. If the first link gain corresponding to a certain AP is greater than the preset link gain transmission threshold... If the number of APs is greater than 1 and less than the total number of APs, the CPU forms a set of candidate APs and, based on the current preset link gain transmission threshold and the current number of relay APs, determines which relay APs from the set will participate in the covert communication between Alice and Bob. If the number of candidate APs does not meet the above conditions, the covert information transmission will not be initiated in the current transmission time slot, or the subsequent parameter optimization process will be entered to adjust the preset link gain transmission threshold. .

[0021] For the determined first Each relay AP, the CPU determines based on the first link CSI, the second link CSI, and a predetermined energy threshold. The relay-assisted channel inversion power control (CIPC) strategy is used to determine the Alice transmit power and the relay AP forwarding power. Specifically, when Alice transmits to the... When a relay AP sends covert information, according to Alice and the first The first link gain between the relay APs determines the transmit power or power allocation factor on the Alice side, making the first... The energy of the signal received at each relay AP reaches a predetermined energy threshold. As shown in equation (1): (1) in, Indicates the predetermined energy threshold. Indicates assignment to the first The power allocation factor of each relay AP, This indicates Alice's transmission power. Indicates Alice and the first Channel coefficients between relay APs Indicates Alice and the first The first link gain between the relay APs.

[0022] In the When a relay AP forwards covert information to Bob, according to the first... The second link gain between the relay AP and Bob is used to determine the forwarding power of the relay AP, enabling Bob to access the relay AP from the second link. The energy of the signal received by the relay AP reaches a predetermined energy threshold. As shown in equation (2): (2) in, Indicates the first The forwarding power of each relay AP, Indicates the first The channel coefficients between the relay AP and Bob. Indicates the first The second link gain between the relay AP and Bob.

[0023] Through the aforementioned CIPC strategy, Alice's transmit power and the relay AP's forwarding power are adaptively adjusted under different link states, ensuring that the received signal energy from Alice to the relay AP, and from the relay AP to Bob, reaches a predetermined energy threshold. Therefore, both Alice's and Bob's sides conceal their communication locations using relay assistance, reducing the likelihood that Eve can infer the locations of the communicating parties based on changes in transmit power, receive power, or link strength.

[0024] After identifying the relay AP, the CPU uses the first link CSI and a preset link gain transmission threshold. The transmission conditions of the current transmission time slot are determined, and the probability of Alice transmitting covert information is determined by combining the statistical parameters of the first link gain. Specifically, whether Alice initiates covert information transmission is determined by the number of APs in the current transmission time slot that meet the first link gain threshold condition; when the number of APs that meet the transmission condition is at least two but less than the total number of APs, the current transmission time slot is considered to meet the probabilistic transmission condition. The same statistical parameters are used for each first link. In the implementation method, the probability of Alice initiating covert information transmission is shown in equation (3): (3) in, This represents the probability that Alice initiates a covert information transmission. This indicates the number of APs that meet the transmission conditions. Indicates the total number of APs. This represents the statistical parameters corresponding to the first link gain. This represents the natural constant. In different implementations of the first link statistical parameters, the probability that an AP meets the preset link gain transmission threshold can be determined based on the first link statistical parameters corresponding to each AP. By enumerating the AP combinations that meet the transmission conditions and accumulating the corresponding combination probabilities, the probability that Alice initiates covert information transmission can be determined.

[0025] When Alice initiates a covert message transmission, the first The received signal model at each relay AP is shown in equation (4): (4) in, Indicates the first The relay AP in the first Signals received when using the secondary channel. This indicates a covert signal sent by Alice. Indicates the first The noise at each relay AP. Equation (4) is used to explain that in the first hop link, the AP received signal consists of the signal component of the covert signal sent by Alice after propagation through the first link and the AP's local noise.

[0026] Because equation (1) makes the first The energy of the signal received at each relay AP reaches a predetermined energy threshold. , No. The signal-to-noise ratio at each relay AP can be related to a predetermined energy threshold. The association is shown in equation (5): (5) in, Indicates the first Signal-to-noise ratio at each relay AP Indicates the first The noise power at each relay AP. Equation (5) is only used to illustrate the correspondence between the CIPC strategy and the first-hop reliable reception in this embodiment.

[0027] After each relay AP decodes the covert information sent by Alice, each relay AP forwards the covert information to Bob according to the forwarding power determined by equation (2). The equivalent receiving model at Bob's location, after taking into account Eve's full-duplex FD interference probability, is shown in equation (6): (6) in, This means Bob is in the Signals received when using the secondary channel. This represents the probability that Eve is in full-duplex jamming / eavesdropping mode. This represents Eve's artificial noise interference power. This represents the interference link channel coefficient from Eve to Bob. This indicates the interference signal sent by Eve. Let represent the local noise at Bob's location. The Eve interference term in equation (6) characterizes the equivalent interference impact calculated based on the probability that Eve is in full-duplex jamming eavesdropping mode, and its corresponding average interference power is: .

[0028] This embodiment establishes a preset Eve mode switching model. The preset Eve mode switching model uses CSI statistics and mode switching thresholds from the Alice to Eve eavesdropping link. The input is used to characterize the probability that Eve selects either half-duplex HD eavesdropping mode or full-duplex jamming eavesdropping mode under different eavesdropping link states. Specifically, when the Alice-Eve eavesdropping link gain is greater than the mode switching threshold... When Eve switches to half-duplex eavesdropping mode and uses two antennas to eavesdrop; when the gain of the Alice-Eve eavesdropping link is less than the mode switching threshold... When Eve switches to full-duplex jamming eavesdropping mode, she uses one antenna to transmit artificial noise and the other antenna to eavesdrop. The probability of Eve being in full-duplex jamming eavesdropping mode is shown in equation (7): (7) in, Indicates the probability of an event occurring. This represents the channel coefficient of the Alice-Eve eavesdropping link. This represents the link gain of the Alice-Eve eavesdropping link. This represents the statistical parameters corresponding to the gain of the eavesdropping link. This indicates the mode switching threshold set by Eve.

[0029] The CPU calls a pre-established, preset concealment constraint evaluation model. Based on the current transmission parameters, the probability of Alice transmitting concealed information, the probability of Eve being in full-duplex interference eavesdropping mode, and the eavesdropping link statistics (CSI), it determines the average minimum error detection probability in full-duplex interference eavesdropping mode and half-duplex eavesdropping mode, and determines the concealment evaluation value corresponding to the current transmission parameters accordingly.

[0030] In the pre-defined hidden constraint evaluation model, consider This indicates that Alice did not transmit any hidden information. This indicates that Alice transmitted a hidden message. Eve uses the received information to perform hypothesis testing to determine whether Alice sent the hidden message.

[0031] For Eve's full-duplex interference eavesdropping mode, the pre-set concealment constraint evaluation model incorporates the probability of Eve being in this mode into the detection power evaluation; under the two assumptions, the equivalent average received power corresponding to the full-duplex interference eavesdropping mode is shown in equation (8): (8) in, This represents the equivalent average power received by Eve in full-duplex jamming eavesdropping mode. This represents Eve's full-duplex self-interference residual coefficient. This represents Eve's self-interference channel coefficient. This represents Eve's self-interference link gain. This represents the noise power at Eve.

[0032] Eve based on detection thresholds right The judgment is made as shown in equation (9): (9) in, This represents Eve's detection threshold in full-duplex jamming eavesdropping mode. This indicates that Eve determined Alice transmitted hidden information. This indicates that Eve determined Alice did not transmit any hidden information.

[0033] The false detection probability of Eve in full-duplex jamming eavesdropping mode consists of the false alarm probability and the false detection probability, as shown in Equation (10): (10) in, This represents the probability of Eve's false detection in full-duplex jamming eavesdropping mode. Indicates the probability of a missed detection. Let represent the false alarm probability. By optimizing the detection thresholds corresponding to equations (8) to (10), the minimum false detection probability of Eve in full-duplex interference eavesdropping mode can be obtained. .

[0034] For Eve's half-duplex eavesdropping mode, the pre-defined concealment constraint evaluation model incorporates the probability of Eve being in this mode into the detection power evaluation; under the two assumptions, the equivalent average received power corresponding to the half-duplex eavesdropping mode is shown in equation (11): (11) in, This represents the equivalent average power received by Eve in half-duplex eavesdropping mode. This indicates Alice to Eve's number of... The channel coefficients corresponding to each eavesdropping antenna This indicates Alice to Eve's number of... The link gain corresponding to each eavesdropping antenna This represents Eve's eavesdropping antenna number. By optimizing the detection threshold corresponding to equation (11), the minimum false detection probability of Eve in half-duplex eavesdropping mode can be obtained. .

[0035] In one implementation, the average minimum error detection probability is obtained by calculating the statistical expectation of the random link gain corresponding to the CSI of the eavesdropping link, as shown in Equation (12): (12) in, This represents Eve's average minimum false detection probability in full-duplex jamming eavesdropping mode. This represents Eve's average minimum false detection probability in half-duplex eavesdropping mode. This represents the statistical expectation. The statistical expectation is based on the eavesdropping link statistics (CSI) and the random link gain used in the detection power evaluation.

[0036] The preset concealment constraint evaluation model takes the smaller of the average minimum error detection probability in the full-duplex interference eavesdropping mode and the half-duplex eavesdropping mode as the concealment evaluation value corresponding to the current transmission parameters, as shown in Equation (13): (13) in, This represents the stealth evaluation value corresponding to the current transmission parameters. The CPU is based on... Determine whether the current transmission parameters satisfy the concealment constraint, as shown in equation (14): (14) in, This represents a positive number that satisfies the hidden constraint requirement.

[0037] The CPU also determines whether Bob can reliably receive covert information based on Bob's reliability constraints. Since Eve applies artificial noise to Bob's legitimate link in full-duplex jamming eavesdropping mode, this embodiment includes the average interference power generated by Eve in full-duplex jamming eavesdropping mode in Bob's reliability constraints. A predetermined energy threshold is also included. The relay AP decoding reliability and Bob reception reliability requirements must be met simultaneously, as shown in equation (15): (15) in, This represents the signal-to-noise ratio threshold required for reliable reception. This represents the noise power at Bob's location. This represents the link gain of the Eve to Bob link. When the noise power of multiple relay APs is different, the first term in equation (15) can be judged separately for each relay AP, or the maximum value among the noise power of each relay AP can be used for conservative judgment.

[0038] like Figure 2 As shown, after acquiring the first link CSI, the second link CSI, and the eavesdropping link statistical CSI, the CPU filters candidate APs based on the first link CSI. When the number of candidate APs meets the probability transmission condition, the relay AP set is determined, and relay-assisted CIPC is used to determine Alice's transmit power and the relay AP's forwarding power. The CPU further determines the probability of Alice transmitting covert information and the probability of Eve being in full-duplex interference eavesdropping mode, and evaluates whether the current transmission parameters meet the covert constraints and Bob's reliability constraints. If the constraints are not met, the parameters to be optimized are updated, and the candidate AP screening and constraint evaluation are performed again. If all constraints are met, the parameter combination used for covert communication is determined, so that Alice transmits and recovers the covert information to Bob via the relay AP.

[0039] Under the condition that both the concealment constraint and the Bob reliability constraint are satisfied, the CPU negotiates and sets a predetermined energy threshold with the legitimate user and the AP. As a power control benchmark, a preset link gain transmission threshold is used. Number of relay APs and predetermined energy threshold The parameters to be optimized are used for joint optimization. The optimization objective is shown in equation (16): (16) in, This indicates Bob's stealth rate. The CPU operates at a fixed predetermined energy threshold. Under these conditions, a preset link gain transmission threshold is determined based on the candidate AP screening results. and the number of relay APs The value of ; within the already determined and Based on this, the predetermined energy threshold is updated using gradient descent or alternating optimization methods. After each update, constraint judgment is performed based on equations (14) and (15), and the parameter combination used for covert communication is updated in the parameter values ​​that satisfy the concealment constraint and Bob reliability constraint, until Bob's concealment rate converges or the preset iteration stop condition is reached.

[0040] When the current transmission time slot meets the probabilistic transmission condition, and the optimized... , and When both the concealment constraint and Bob's reliability constraint are satisfied, Alice sends concealed information to each relay AP at a determined transmit power. Each relay AP receives and decodes the concealed information sent by Alice, and then forwards it to Bob at a determined forwarding power. Bob receives the forwarded signals from multiple relay APs and recovers the concealed information, thus completing the concealed communication between Alice and Bob.

[0041] In an alternative implementation, each wireless link is modeled as an independent Rayleigh fading channel, with the channel gain remaining constant within the same transmission time slot and varying independently between different transmission time slots. Based on this channel model, the Alice-Eve link gain, Alice-AP link gain, and AP-Bob link gain can be statistically described using an exponential distribution. This Rayleigh fading and exponential distribution description is used to support the probabilistic transmission calculation in Equation (3) and the Eve mode switching probability calculation in Equation (7), and is not considered a technical feature independent of the steps of this method.

[0042] In another alternative implementation, Bob's concealment rate The equivalent signal-to-interference-plus-noise ratio (SINNR) at Bob's end can be determined. Bob's equivalent SINNR is determined by the effective signal components forwarded to Bob by multiple relay APs, the noise components at Bob's end, and the average interference components generated when Eve is in full-duplex jamming eavesdropping mode. This equivalent SINNR is used to evaluate Bob's concealment rate in Equation (16) and is consistent with Bob's reliability constraints in Equation (15).

[0043] In another alternative implementation, under normalized bandwidth, the equivalent signal-to-interference-plus-noise ratio (SINNR) at the Bob end can be substituted into the logarithmic rate evaluation function to obtain the Bob concealment rate. This rate evaluation function is used for different... , and The combinations are compared to select the parameter combination that improves Bob's concealment rate.

[0044] Example 2 like Figure 3 As shown, this embodiment provides a non-cellular IoT covert communication system applied to a non-cellular IoT network where a smart eavesdropper named Eve exists. The system includes a central processing unit (CPU), multiple access points (APs), and multiple IoT devices. Any one of the IoT devices can be configured by the CPU as either a transmitter (Alice) or a receiver (Bob) within a transmission time slot. The multiple APs communicate with the CPU via backhaul links to assist Alice and Bob in completing covert communication. The smart eavesdropper Eve, as an external eavesdropping entity, represents the smart eavesdropping environment faced by this system.

[0045] The CPU comprises a scheduling control module, a CSI management module, a candidate AP screening module, a relay-assisted CIPC power control module, a probabilistic transmission judgment module, an Eve mode evaluation module, a concealment constraint evaluation module, a Bob reliability evaluation module, a parameter optimization module, and a transmission control module. Each module is implemented using a CPU-side processor, memory, and a backhaul communication interface, and interacts with other modules through the CPU's internal control logic. The CPU sends scheduling commands, channel estimation commands, power control parameters, and forwarding control commands to the APs via the backhaul link, and receives link CSI feedback from the APs, as well as execution information related to channel estimation and relay forwarding.

[0046] Each access point (AP) includes an AP communication interface, a channel estimation unit, a CSI feedback unit, a covert information receiving unit, a decoding unit, and a relay forwarding unit. The AP communication interface is used for backhaul communication with the CPU and for wireless communication with IoT devices. The channel estimation unit estimates the first link CSI between Alice and the AP, and the second link CSI between the AP and Bob, based on the pilot signals within the transmission time slot, after the CPU issues a channel estimation command. The CSI feedback unit feeds back the first and second link CSIs to the CPU; the first link CSI, used to determine Alice's transmit power or power allocation factor, can be fed back to Alice by the CSI feedback unit or issued to Alice by the CPU based on the feedback result. The covert information receiving unit receives covert information sent by Alice, the decoding unit decodes the covert information, and the relay forwarding unit forwards the covert information to Bob according to the forwarding power determined by the CPU.

[0047] Each IoT device includes a device communication interface and a covert communication execution unit. The device communication interface receives role scheduling information and power control information from the CPU. When the IoT device is configured as Alice, the covert communication execution unit sends covert information to the relay AP according to the Alice transmit power determined by the CPU; when the IoT device is configured as Bob, it receives the forwarded signal from the relay AP and restores the covert information. Unscheduled IoT devices remain silent during the current transmission time slot.

[0048] Multiple transmission time slots are set within a transmission cycle. For any given transmission time slot, the scheduling control module schedules a pair of IoT devices from multiple IoT devices, identifying them as Alice and Bob, and notifies each AP to enter the channel estimation state. Each AP, triggered by the channel estimation command, estimates the first link CSI and the second link CSI based on pilot signals and feeds them back to the CSI management module via the backhaul link.

[0049] The CSI management module receives and organizes the first link CSI and the second link CSI. It then sends the first link CSI to the candidate AP filtering module and both the first and second link CSIs to the relay auxiliary CIPC power control module. The candidate AP filtering module obtains a preset link gain transmission threshold. The first link gain between Alice and each AP is compared with the threshold. If the first link gain of an AP is greater than the threshold, the AP is identified as a candidate AP. If the number of candidate APs is greater than 1 and less than the total number of APs, the candidate AP screening module sends the candidate AP set to the probability transmission judgment module and the parameter optimization module for subsequent determination of the number of relay APs and the relay AP set. If the number of candidate APs does not meet the above conditions, the current transmission time slot will not initiate covert information transmission, or the parameter optimization module will update the preset link gain transmission threshold and re-determine the candidate APs in the subsequent screening process.

[0050] The relay-assisted CIPC power control module is used to control the power of relays based on the received relay AP set and a predetermined energy threshold. Then, based on the first link CSI, the second link CSI, the relay AP set, and the predetermined energy threshold, the Alice transmit power and the forwarding power of each relay AP are determined. Specifically, this module determines the Alice-side transmit power or power allocation factor according to formula (1) in Example 1, so that the first link CSI, the second link CSI, the relay AP set, and the predetermined energy threshold are all used to determine the Alice transmit power and the forwarding power of each relay AP. The energy of the signal received at each relay AP reaches a predetermined energy threshold; and the energy of the signal received at each relay AP is determined according to formula (2) in Example 1. The forwarding power of the relay AP is adjusted so that the energy of the signal Bob receives from the relay AP reaches a predetermined energy threshold.

[0051] The probability transmission determination module determines the probability of Alice transmitting covert information based on the first link CSI, the preset link gain transmission threshold, and the number of APs that meet the threshold conditions. When each first link uses the same statistical parameters, the probability transmission judgment module determines the probability according to formula (3) in Example 1; when the statistical parameters of each first link are different, the probability is determined by enumerating the AP combinations that meet the transmission conditions and accumulating the corresponding combination probabilities.

[0052] The Eve mode evaluation module calls the Eve mode switching model pre-stored in the CPU-side memory. This model uses pre-configured Alice-Eve eavesdropping link statistics (CSI) and mode switching thresholds. The input is used to characterize the probability of Eve switching between half-duplex eavesdropping mode and full-duplex jamming eavesdropping mode. The Eve mode evaluation module determines the probability of Eve being in full-duplex jamming eavesdropping mode according to equation (7) in Example 1. .

[0053] The concealment constraint evaluation module calls the preset concealment constraint evaluation model and evaluates whether the current transmission parameters meet the concealment constraints based on the current transmission parameters, the probability of Alice transmitting concealed information, the probability of Eve being in full-duplex interference eavesdropping mode, and the eavesdropping link statistics (CSI). The concealment constraint evaluation module determines the concealment evaluation value corresponding to the current transmission parameters based on the FD / HD detection power evaluation, detection threshold optimization, and average minimum error detection probability calculation process corresponding to equations (8) to (13) in Example 1, and judges whether the current transmission parameters meet the concealment constraints based on equation (14) in Example 1.

[0054] The Bob reliability evaluation module determines whether Bob can reliably receive covert information based on a predetermined energy threshold, the set of relay APs, the noise power at each relay AP, the noise power at Bob's location, and the average interference power generated when Eve is in full-duplex interference eavesdropping mode. The Bob reliability evaluation module determines whether the predetermined energy threshold simultaneously satisfies the relay AP decoding reliability and Bob's reception reliability based on formula (15) in Example 1. When the noise power of multiple relay APs differs, each relay AP is evaluated separately, or the maximum value among the noise powers of each relay AP is used for a conservative evaluation.

[0055] The parameter optimization module receives the candidate AP set, the probability of Alice transmitting covert information, and the probability of Eve being in full-duplex interference eavesdropping mode. It then jointly optimizes the preset link gain transmission threshold, the number of relay APs, and a predetermined energy threshold, updating the values ​​of these parameters during the optimization process. The parameter optimization module calls the covert constraint evaluation module and the Bob reliability evaluation module to obtain constraint judgment results, and determines the parameter combination used for covert communication from the parameter values ​​that satisfy the covert constraints and Bob reliability constraints. After determining the number of relay APs, the parameter optimization module selects the corresponding relay AP set from the candidate AP set and outputs it to the relay-assisted CIPC power control module and the transmission control module.

[0056] Specifically, the parameter optimization module performs the optimization objective according to formula (16) in Example 1, and optimizes the preset link gain transmission threshold. Number of relay APs and predetermined energy threshold Joint optimization is performed. The parameter optimization module first fixes a predetermined energy threshold. Determine the preset link gain transmission threshold from the candidate AP screening results. and the number of relay APs The value of ; then, within the predetermined preset link gain transmission threshold. and the number of relay APs Based on this, the predetermined energy threshold is updated using gradient descent or alternating optimization methods. After each update, a concealment constraint evaluation and a Bob reliability constraint evaluation are performed, and the parameter combination used for covert communication is updated among the parameter values ​​that satisfy the concealment constraint and the Bob reliability constraint, until the Bob concealment rate converges or the preset iteration stopping condition is reached.

[0057] After determining the parameter combination for covert communication, the parameter optimization module sends the determined predetermined energy threshold and relay AP set to the relay-assisted CIPC power control module. Based on the determined predetermined energy threshold and relay AP set, the relay-assisted CIPC power control module determines Alice's transmit power and the forwarding power of each relay AP, and sends the power control parameters to the transmission control module.

[0058] The transmission control module generates transmission control commands based on the relay AP set output by the parameter optimization module and the Alice transmit power and relay AP forwarding power output by the relay auxiliary CIPC power control module. The transmission control module sends the power control parameters used by Alice to transmit covert information to Alice, sends the relay AP set or corresponding relay AP identifiers to Alice, sends the forwarding power of each relay AP to the corresponding relay AP, and issues forwarding control commands to each relay AP.

[0059] When the current transmission time slot meets the above-mentioned candidate AP quantity condition, and the determined transmission parameters simultaneously meet the concealment constraint and Bob's reliability constraint, Alice sends concealment information to each relay AP according to the transmit power determined by the CPU. After receiving and decoding the concealment information, each relay AP forwards the concealment information to Bob according to the forwarding power determined by the CPU. Bob receives the forwarded signals from multiple relay APs and recovers the concealment information.

[0060] Therefore, the system forms the following data flow: The CPU sends scheduling and channel estimation instructions to the APs and IoT devices; the APs feed back the first link CSI and the second link CSI to the CPU; the CPU filters the candidate AP set based on the first link CSI and the preset link gain transmission threshold; the CPU determines the parameter combination for covert communication based on the probability of Alice transmitting covert information, the probability of Eve being in full-duplex interference eavesdropping mode, the covert constraint evaluation results, and the Bob reliability evaluation results, and determines the relay AP set from the candidate AP set; the CPU determines Alice's transmit power and the relay AP's forwarding power based on the determined relay AP set, the first link CSI, the second link CSI, and the preset energy threshold; the CPU sends the transmission parameters and power control parameters to Alice and the corresponding relay AP; Alice sends covert information to the relay AP, and the relay AP forwards the covert information to Bob.

[0061] Through the above system architecture, the CPU centrally handles scheduling, CSI management, AP selection, power control, concealment constraint evaluation, reliability evaluation, and parameter optimization. The APs perform link CSI estimation, feedback, concealed information reception, and relay forwarding. IoT devices dynamically assume the roles of Alice or Bob based on CPU scheduling. Therefore, this system can achieve probabilistic concealed transmission based on legitimate link CSI in non-cellular IoT networks where the intelligent eavesdropper Eve exists. Furthermore, through a relay-assisted CIPC strategy, the received signal energy at both the relay AP and Bob reaches a predetermined energy threshold, while simultaneously improving Bob's concealment rate while satisfying both concealment constraints and Bob's reliability constraints.

Claims

1. A covert communication method for non-cellular IoT, applied to a non-cellular IoT network where a smart eavesdropper named Eve exists, characterized in that, include: The central processing unit (CPU) schedules a pair of IoT devices as Alice, the transmitter, and Bob, the receiver, and notifies multiple access points (APs) to enter the channel estimation state. Obtain the first link channel status information (CSI) between Alice and each AP, and the second link CSI between each AP and Bob; Based on the first link CSI and the preset link gain transmission threshold, candidate APs that meet the transmission conditions are selected from multiple APs to form a candidate AP set, and the probability of Alice transmitting covert information is determined. Based on the CSI statistics and mode switching threshold of the Alice-Eve eavesdropping link, the probability of Eve switching between half-duplex eavesdropping mode (HD) and full-duplex jamming eavesdropping mode (FD) is determined. The concealment constraint is determined based on the average minimum error detection probability of Eve in HD and FD. The reliability constraint of Bob is determined based on the average interference power generated when Eve is in FD. The preset link gain transmission threshold, the number of relay APs, and the predetermined energy threshold are jointly optimized, and the parameter combination used to optimize the Bob concealment rate is determined from the preset link gain transmission threshold, the number of relay APs, and the predetermined energy threshold values ​​that satisfy the concealment constraint and the Bob reliability constraint. The relay AP set is determined from the candidate AP set based on the parameter combination, and the relay auxiliary channel inversion power control (CIPC) strategy is used to determine Alice's transmit power and the relay AP's forwarding power based on the first link CSI, the second link CSI and the predetermined energy threshold, so that the received signals from Alice to the relay AP and from the relay AP to Bob both reach the predetermined energy threshold. Based on Alice's transmit power, relay AP forwarding power, and the relay AP set, relay-assisted covert communication is performed between Alice and Bob.

2. The non-cellular IoT covert communication method according to claim 1, characterized in that, Based on the first link CSI and a preset link gain transmission threshold, candidate APs that meet the transmission conditions are selected from multiple APs to form a candidate AP set, including: The first link gain between Alice and each AP is compared with the preset link gain transmission threshold. If the first link gain corresponding to a certain AP is greater than the preset link gain transmission threshold, then the corresponding AP is determined as a candidate AP; If the number of candidate APs is greater than 1 and less than the total number of APs, then the candidate APs are combined into a candidate AP set. If the number of candidate APs does not meet the condition of being greater than 1 and less than the total number of APs, then the current transmission time slot will not initiate covert information transmission, or it will enter the subsequent parameter optimization process to adjust the preset link gain transmission threshold.

3. The non-cellular IoT covert communication method according to claim 1, characterized in that, The relay-assisted CIPC strategy is used to determine the Alice transmit power and the relay AP forwarding power, including: Based on the first link gain between Alice and any relay AP, the predetermined energy threshold, and the power allocation factor allocated to any relay AP, determine the transmit power of Alice so that the energy of the received signal at any relay AP reaches the predetermined energy threshold. Based on the second link gain between any relay AP and Bob and a predetermined energy threshold, determine the forwarding power of any relay AP so that the energy of the signal Bob receives from any relay AP reaches the predetermined energy threshold. The energy of the signal received at any relay AP is determined by Alice's transmit power, power allocation factor, and corresponding first link gain. The energy of the signal received by Bob from any relay AP is determined by the forwarding power of any relay AP and corresponding second link gain.

4. The non-cellular IoT covert communication method according to claim 1, characterized in that, Based on the first link CSI and the preset link gain transmission threshold, determine the probability of Alice transmitting covert information, including: Based on the number of APs in the current transmission time slot whose first link gain is greater than the preset link gain transmission threshold, determine whether the current transmission time slot meets the probabilistic transmission condition; When the number of APs whose first link gain is greater than the preset link gain transmission threshold is at least two and less than the total number of APs, the current transmission time slot is determined to meet the probabilistic transmission condition. When all first links use the same statistical parameters, the probability of Alice initiating covert information transmission is determined based on the total number of APs, the statistical parameters corresponding to the first link gain, and the preset link gain transmission threshold. When different statistical parameters are used for each first link, the probability of each AP meeting the preset link gain transmission threshold is determined. By enumerating the AP combinations that meet the transmission conditions and accumulating the corresponding combination probabilities, the probability of Alice initiating covert information transmission is determined.

5. The non-cellular IoT covert communication method according to claim 1, characterized in that, Determine the probability of Eve switching between HD and FD, including: When the gain of the Alice-Eve eavesdropping link is greater than the mode switching threshold, it is determined that Eve has switched to HD and is using two antennas to eavesdrop. When the gain of the Alice-Eve eavesdropping link is less than the mode switching threshold, it is determined that Eve has switched to FD and uses one antenna to send artificial noise and the other antenna to eavesdrop. Based on the CSI statistics and mode switching threshold of the Alice-Eve eavesdropping link, determine the probability that Eve is in FD.

6. The non-cellular IoT covert communication method according to claim 1, characterized in that, The concealment constraints are determined based on Eve's average minimum false detection probability under HD and FD, including: Based on the current transmission parameters, the probability of Alice transmitting covert information, the probability of Eve being in FD, and the eavesdropping link statistics CSI, determine Eve's average minimum error detection probability under FD and Eve's average minimum error detection probability under HD respectively. The smaller of Eve's average minimum false detection probability under FD and Eve's average minimum false detection probability under HD is used as the concealment evaluation value; Based on the comparison between the concealment evaluation value and the concealment constraint value, it is determined whether the current transmission parameters meet the concealment constraints. The concealment constraint value is determined based on the smaller of the probability that Alice will transmit concealed information and the probability that Alice will not transmit concealed information, and a preset positive number.

7. The non-cellular IoT covert communication method according to claim 1, characterized in that, Determine Bob's reliability constraints, including: Based on the predetermined energy threshold, the set of relay APs, the noise power at each relay AP, the noise power at Bob, and the average interference power generated when Eve is in FD, determine whether the predetermined energy threshold simultaneously satisfies the decoding reliability of the relay APs and the receiving reliability of Bob. When the noise power of multiple relay APs is different, each relay AP should be judged separately, or the maximum value among the noise power of each relay AP should be used for conservative judgment.

8. The non-cellular IoT covert communication method according to claim 1, characterized in that, The preset link gain transmission threshold, the number of relay APs, and the predetermined energy threshold are jointly optimized. Within the preset link gain transmission threshold, the number of relay APs, and the predetermined energy threshold, which satisfy both concealment constraints and Bob's reliability constraints, the parameter combinations used to optimize Bob's concealment rate are determined, including: The preset link gain transmission threshold, the number of relay APs, and the predetermined energy threshold are used as parameters to be optimized. Under the condition of a fixed predetermined energy threshold, the values ​​of the preset link gain transmission threshold and the number of relay APs are determined based on the candidate AP screening results and Bob's concealment rate. Based on the predetermined link gain transmission threshold and the number of relay APs, the predetermined energy threshold is updated using gradient descent or alternating optimization methods. After each update of the predetermined energy threshold, a concealment constraint evaluation and a Bob reliability constraint evaluation are performed. The parameter combination used for covert communication is updated based on the values ​​of the preset link gain transmission threshold, the number of relay APs, and the predetermined energy threshold that satisfy the concealment constraint and the Bob reliability constraint, until the Bob concealment rate converges or the preset iteration stopping condition is reached.

9. A non-cellular IoT covert communication system, applied to a non-cellular IoT network where a smart eavesdropper named Eve exists, characterized in that, This includes a central processing unit (CPU), multiple access points (APs), and multiple IoT devices; The CPU is used to schedule a pair of IoT devices as Alice and Bob, notify multiple APs to enter the channel estimation state, obtain the first link CSI between Alice and each AP and the second link CSI between each AP and Bob, filter candidate APs based on the first link CSI and the preset link gain transmission threshold to form a candidate AP set, determine the probability of Alice transmitting covert information, determine the probability of Eve switching between half-duplex eavesdropping mode (HD) and full-duplex interference eavesdropping mode (FD) based on the Alice-Eve eavesdropping link statistical CSI and mode switching threshold, determine the covert constraints based on the average minimum error detection probability of Eve in HD and FD, determine the Bob reliability constraints based on the average interference power generated by Eve in FD, jointly optimize the preset link gain transmission threshold, the number of relay APs and the predetermined energy threshold, and determine the parameter combination used to optimize Bob's covert rate from the values ​​of the preset link gain transmission threshold, the number of relay APs and the predetermined energy threshold that satisfy the covert constraints and Bob reliability constraints, determine the relay AP set from the candidate AP set based on the parameter combination, and use the relay-assisted CIPC strategy to determine Alice's transmit power and the relay AP forwarding power. The AP is used to perform link CSI estimation and feedback under CPU control, and when it is identified as a relay AP, it receives covert information sent by Alice, decodes the covert information, and forwards the covert information to Bob. The IoT device is used to send covert information according to the Alice transmit power determined by the CPU when it is configured as Alice by the CPU, and to receive the forwarded signal from the relay AP and restore the covert information when it is configured as Bob by the CPU.

10. A non-cellular IoT covert communication system according to claim 9, characterized in that, The CPU includes a scheduling control module, a CSI management module, a candidate AP screening module, a relay-assisted CIPC power control module, a probability transmission judgment module, an Eve mode evaluation module, a hidden constraint evaluation module, a Bob reliability evaluation module, a parameter optimization module, and a transmission control module. The scheduling control module is used to schedule Alice and Bob and notify the AP to enter the channel estimation state; The CSI management module is used to receive and organize the first link CSI and the second link CSI. The candidate AP filtering module is used to filter candidate APs based on the first link CSI and a preset link gain transmission threshold. The relay-assisted CIPC power control module is used to determine the Alice transmit power and the forwarding power of each relay AP; The probability transmission judgment module is used to determine the probability that Alice will transmit covert information; The Eve mode evaluation module is used to determine the probability that Eve is in FD based on the CSI statistics and mode switching threshold of the Alice-Eve eavesdropping link. The concealment constraint evaluation module is used to evaluate whether the values ​​of the preset link gain transmission threshold, the number of relay APs, and the predetermined energy threshold satisfy the concealment constraints based on Eve's average minimum error detection probability under FD and HD. The Bob reliability evaluation module is used to evaluate whether the values ​​of the preset link gain transmission threshold, the number of relay APs, and the predetermined energy threshold meet the Bob reliability constraints based on the average interference power generated when Eve is in FD. The parameter optimization module is used to jointly optimize the preset link gain transmission threshold, the number of relay APs, and the predetermined energy threshold, and to determine the parameter combination for covert communication from the preset link gain transmission threshold, the number of relay APs, and the predetermined energy threshold values ​​that satisfy the concealment constraint and the Bob reliability constraint. The transmission control module is used to send power control parameters and the set of relay APs to Alice and the corresponding relay APs, so as to control Alice to send covert information to the relay APs and control the relay APs to forward covert information to Bob.

Images