CN-121985335-A - Method, device, equipment and medium for unloading computing task security integrating communication and perception

Abstract

The invention relates to the technical field of wireless communication and network security and discloses a method, a device, equipment and a medium for safely unloading calculation tasks, wherein the method comprises the steps of transmitting a composite signal, receiving echo signals corresponding to sensing signals, wherein the composite signal comprises the communication signals and the sensing signals, estimating an eavesdropping position according to the echo signals, wherein the eavesdropping position comprises an angle uncertainty interval of an eavesdropper, inputting system state information and the eavesdropping position into a decision model, outputting a control strategy through the decision model, and adjusting the direction and energy of the composite signal by the control strategy so as to strengthen the communication signals received by a receiving end and weaken other signals received by the eavesdropper, and adjusting the composite signal according to the control strategy so as to unload the tasks to the receiving end.

Inventors

• LIANG HUI
• WU ZHIHUI
• ZHANG GUOBIN

Assignees

• 东莞理工学院

Dates

Publication Date

20260505

Application Date

20251211

Claims (10)

1. 1. A method for integrating communication and perceived computational task security offload, characterized in that it is applied to a sender, said method comprising: Transmitting a composite signal and receiving a corresponding echo signal, wherein the composite signal comprises a communication signal and a sensing signal, and the echo signal is the echo signal corresponding to the sensing signal; Estimating an eavesdropping position of an eavesdropper according to the echo signals, wherein the eavesdropping position comprises altitude information and direction information; Inputting system state information and the eavesdropping position into a decision model, and outputting a control strategy through the decision model, wherein the system state information is state information of a communication network, the communication network is a communication network of a transmitting end and a receiving end, the control strategy is used for adjusting the direction and energy of the composite signal so as to enhance the communication signal received by the receiving end and weaken other received signals, and enhance the perception signal received by the eavesdropper and weaken the received communication signal; and adjusting the composite signal according to the control strategy so as to safely unload the task to the receiving end.
2. 2. The method of claim 1, wherein the system state information comprises channel state information of the transmitting end to the receiving end, signal-to-interference-plus-noise ratio of a current communication link, signal-to-interference-plus-noise ratio of a current perceived link, and instantaneous power consumption of the transmitting end.
3. 3. The method of claim 1, wherein the decision model is a deep reinforcement learning model, the inputting system state information and the tap location into the decision model, and outputting a control strategy through the decision model, comprising: Inputting the system state information and the eavesdropping position as the reinforcement learning environment information into the decision model, and outputting a communication covariance matrix and a perception covariance matrix serving as an action space through the decision model; Generating a first beamforming vector for controlling the communication signal according to the communication covariance matrix; and generating a second beam forming vector for controlling the sensing signal according to the sensing covariance matrix.
4. 4. The method of claim 3, wherein when there are a plurality of the transmitting ends, the decision model outputs the communication covariance matrix and the sensing covariance matrix according to the system state information and the eavesdropping position, comprising: receiving system state information and interception positions which are correspondingly transmitted by each transmitting end; Analyzing all the received system state information and all the eavesdropping positions as environment information to obtain a covariance matrix set serving as an action space; And feeding back a target communication covariance matrix and a target perception covariance matrix corresponding to a target sending end in the covariance matrix set to the target sending end.
5. 5. The method of claim 4, wherein said estimating the eavesdropping location of an eavesdropper from the echo signal comprises: estimating reference direction angle and altitude information of the eavesdropper based on the echo signal; Modeling an angle uncertainty section through a perception error and the reference direction angle, and taking the angle uncertainty section as direction information of the eavesdropper.
6. 6. The method of claim 5, wherein the reward function of the decision model comprises at least one of a first scoring term, a second scoring term, a third scoring term, and a fourth scoring term, wherein the first scoring term is used for determining according to a lowest data transmission rate in each sender, the second scoring term is used for determining according to a perceived signal-to-interference ratio of a link of each sender, the third scoring term is used for determining according to a maximum eavesdropping signal-to-interference ratio, the maximum eavesdropping signal-to-interference ratio is a maximum eavesdropping signal-to-interference ratio predicted within an angle uncertainty interval corresponding to each sender, the eavesdropping signal-to-interference ratio is an eavesdropping link signal-to-interference ratio when eavesdropping, and the fourth scoring term is used for determining according to instantaneous power consumption of the sender.
7. 7. The method of claim 4, wherein a sum of a communication covariance matrix and a perceptual covariance matrix corresponding to any transmitting end satisfies a preset maximum transmit power constraint.
8. 8. A computing task security offload device integrating communication and awareness, characterized by a sender, the device comprising: The signal synthesis and transmission module is used for transmitting a composite signal and receiving a corresponding echo signal, wherein the composite signal comprises a communication signal and a perception signal, and the echo signal is the echo signal corresponding to the perception signal; the positioning module is used for estimating the eavesdropping position of the eavesdropper according to the echo signals, and the eavesdropping position comprises height information and direction information; The strategy decision module is used for inputting system state information and the eavesdropping position into a decision model, outputting a control strategy through the decision model, wherein the system state information is state information of a communication network, the communication network is a communication network of a transmitting end and a receiving end, the control strategy is used for adjusting the direction and energy of the composite signal so as to strengthen the communication signal received by the receiving end and weaken other received signals, and strengthen the perception signal received by the eavesdropper and weaken the received communication signal; And the composite signal adjustment and task unloading module is used for adjusting the composite signal according to the control strategy so as to safely unload the task to the receiving end.
9. 9. An electronic device, comprising: A memory and a processor in communication with each other, the memory having stored therein computer instructions which, upon execution, cause the processor to perform the method of any of claims 1 to 7.
10. 10. A computer readable storage medium having stored thereon computer instructions for causing a computer to perform the method of any one of claims 1 to 7.

Description

Method, device, equipment and medium for unloading computing task security integrating communication and perception Technical Field The invention relates to the technical field of wireless communication and network security, in particular to a method, a device, equipment and a medium for unloading computing task security integrating communication and perception. Background In the current wireless communication system, especially the system facing complex scenes such as the internet of things and the internet of vehicles, the network node always needs to have the environment sensing capability while finishing the main communication service so as to realize understanding and response to the surrounding situation. The purpose is to prevent malicious eavesdropping nodes (such as malicious unmanned aerial vehicles) from eavesdropping communication signals sent to a receiving end (such as a satellite) by a sending end (such as an Internet of things device). However, the sensing function (for scanning for eavesdroppers) and the communication function (for transmitting useful communication signals) of the existing secure transmission scheme are often implemented with different devices, respectively, which is disadvantageous for low-cost resources and high communication efficiency. Disclosure of Invention The invention provides a method, a device, equipment and a medium for safely unloading an integrated communication and perception computing task, which are used for solving the problem that the existing wireless communication system lacks self-adaptive protection capability for mobile eavesdropping threat in a dynamic open environment. The invention provides a method for safely unloading calculation tasks integrating communication and perception, which comprises the steps of transmitting a composite signal and receiving corresponding echo signals, wherein the composite signal comprises a communication signal and a perception signal, the echo signals are echo signals corresponding to the perception signal, estimating an eavesdropping position of an eavesdropper according to the echo signals, the eavesdropping position comprises altitude information and direction information, inputting system state information and the eavesdropping position into a decision model, and outputting a control strategy through the decision model, wherein the system state information is state information of a communication network, the communication network is a communication network of a transmitting end and a receiving end, the control strategy is used for adjusting the direction and energy of the composite signal so as to enable the communication signal received by the receiving end to be enhanced and other signals received by the eavesdropper to be weakened, and adjusting the composite signal according to the control strategy so as to safely unload the calculation tasks to the receiving end. In an alternative embodiment, the system state information includes channel state information from sender to receiver, signal to interference plus noise ratio of the current communication link, signal to interference plus noise ratio of the current perceived link, and instantaneous power consumption of the sender. In an alternative implementation mode, the decision model is a deep reinforcement learning model, the system state information and the eavesdropping position are input into the decision model, and a control strategy is output through the decision model, and the method comprises the steps of inputting the system state information and the eavesdropping position serving as environment information of reinforcement learning into the decision model, outputting a communication covariance matrix and a perception covariance matrix serving as action spaces through the decision model, generating a first beam forming vector for controlling communication signals according to the communication covariance matrix, and generating a second beam forming vector for controlling the perception signals according to the perception covariance matrix. In an optional implementation manner, when a plurality of sending ends exist, the decision model outputs a communication covariance matrix and a perception covariance matrix according to the system state information and the interception positions, and the method comprises the steps of receiving the system state information and the interception positions which are correspondingly sent by each sending end, analyzing all the received system state information and all the received interception positions as environment information to obtain a covariance matrix set serving as an action space, and feeding back a target communication covariance matrix and a target perception covariance matrix which correspond to a target sending end in the covariance matrix set to the target sending end. In an alternative embodiment, estimating the eavesdropping location of the eavesdropper from the echo signal includes estimating reference di