A method for preventing privacy leakage in a beamforming process based on a router end space mapping mechanism

Abstract

The application discloses a method for preventing privacy leakage in a beam forming process based on a router end space mapping mechanism, which reuses the space mapping mechanism of the router, confuses the signal used in channel detection at the router end, and sends the confused steering matrix back to the router end after normal channel detection at the receiving end, so that the hardware and processing flow of the receiving end do not need to be modified, the confusion of the signal and the restoration of the steering matrix are completed at the router end, zero modification of the receiving end is realized, and hardware modification is greatly reduced. Compared with the prior art, the application significantly reduces the cost of wide application, has little influence on communication, and is suitable for practical scenarios.

Description

TECHNICAL FIELD

[0001] The present application relates to the problem of privacy leakage in the field of WiFi, in particular, to the problem of privacy leakage in the beamforming technology in the field of WiFi, and is a method for preventing privacy leakage in the beamforming process based on a router-end space mapping mechanism. BACKGROUND

[0002] In recent years, with the rapid increase in the number of and variety of smart devices supporting WiFi, they have been widely deployed. Beamforming technology is an important direction of WiFi technology, which can enhance WiFi signals and increase signal-to-noise ratio. Beamforming technology is a technology that improves the quality of signal transmission in a wireless communication system by weighting the transmitted signals. It can use an antenna array or multiple individual antennas to focus and direct signals to achieve better signal quality at the target device or receiver. It has two main improvements for traditional WiFi: it enhances antenna signals, speeds up transmission, and increases signal range; and it isolates the collision domain in space, increasing network throughput. Therefore, beamforming technology is currently widely used in various WiFi devices.

[0003] However, there is a problem of privacy leakage in the beamforming process. The beamforming process is initiated by the router end, and after the receiving end performs channel measurement, a steering matrix needs to be calculated and sent back to the router end, which uses the steering matrix to guide the spatial mapping of the transmitting antenna. The steering matrix implies the spatial information around the receiving end, although the steering matrix has lost part of the spatial information and only retains the part of the information that guides the steering of the antenna, but after processing, it still leaks the user's privacy information, such as location and activity. Moreover, the transmission of the steering matrix is in clear text, and the attacker can obtain the corresponding information from the data packet without decryption, so the beamforming process has the risk of user location privacy leakage.

[0004] A current solution is to encrypt the beamforming feedback packet CBF. This method can effectively prevent the problem of privacy leakage, but the computational overhead brought by encryption may cause additional burden on communication. In addition, this method needs to update the firmware on all WiFi devices to support encryption, and considering the huge number of existing WiFi devices, it is a very difficult task to widely apply this method. Another method is to inject fake beamforming feedback packets CBF to confuse the information collected by the attacker, but the router may not be able to distinguish between fake beamforming feedback packets CBF, resulting in the use of incorrect steering matrices for beamforming, which seriously reduces the communication performance. There is no good solution to the problem of privacy leakage in the beamforming process. SUMMARY

[0005] In view of the shortcomings of the prior art, the technical problem to be solved by the present invention is to overcome the problems of high computational overhead affecting communication performance and large hardware modifications leading to high cost for widespread application when solving the privacy leakage problem in the beamforming process. The present invention provides a method for preventing privacy leakage in the beamforming process based on the router-side spatial mapping mechanism. That is, the router's spatial mapping mechanism is reused, and the signal used for channel probing is obfuscated at the router. The receiver performs channel probing normally and sends the obfuscated turning matrix back to the router. The hardware and processing flow of the receiver do not need to be modified, ensuring minimal hardware modification. The turning matrix is ​​restored at the router, ensuring that beamforming can be performed normally.

[0006] The present invention has been implemented using the following technical solutions:

[0007] A method for preventing privacy leakage during beamforming based on router-side spatial mapping mechanism includes the following steps:

[0008] S1. The router generates a spatiotemporal confusion matrix Q based on the spatial mapping mechanism;

[0009] S2. Use the spatiotemporal confusion matrix Q obtained in step S1 as the spatial mapping matrix during router beamforming channel detection.

[0010] S3. The router performs spatial mapping and sends an empty data packet (NDP) to the receiving end.

[0011] S4. The receiving end calculates the confused redirection matrix V based on the NDP received in step S3. obf ;

[0012] S5. The receiving end will use the turning matrix V obtained in step S4. obf After quantization and compression, the data is sent back to the router in the form of a compressed beamforming feedback packet (CBF).

[0013] S6. The router decompresses the compressed beamforming feedback packet (CBF) received in step S5 to obtain the confused steering matrix V. obf ;

[0014] S7. The router uses the spatiotemporal confusion matrix Q obtained in step S1 to modify the confused redirection matrix V obtained in step S6. obf Perform the restoration to obtain the restored steering matrix V;

[0015] S8. The router uses the restored steering matrix V for beamforming.

[0016] As a further improvement, in step S1, the router end generates a space-time confusion matrix Q according to a space mapping mechanism, which is used to confuse the signal sent during channel probing, specifically:

[0017] Obtaining the channel bandwidth and the protocol type; referring to a channel bandwidth and protocol type corresponding relationship table, determining the number of subcarriers k of the channel according to the channel bandwidth and the protocol type; using QR decomposition technology to randomly generate a k×k unitary matrix Q, which is used to confuse the signal used during subsequent channel probing.

[0018] As a further improvement, in step S2, the space-time confusion matrix Q obtained in step S1 is used as the space mapping matrix for router beamforming channel probing, which is used to confuse the signal sent during channel probing, specifically:

[0019] The default direct mapping space mapping matrix of the router end transmitting antenna mapping to the space flow is replaced with the confusion matrix Q obtained in step S1 to perform space mapping on the transmitting antenna. The signal sent during channel probing is the long training field LTF information X in the null data packet NDP, and the LTF is a fixed signal in a protocol. The default direct mapping space mapping matrix will send the LTF to the receiving end without any change. After using the confusion matrix Q to perform space mapping on the transmitting antenna, the signal transmitted by the router end to the receiving end is represented as:

[0020] X obf =QX

[0021] Where X obf is the confused signal, Q is the confusion matrix, and X is the LTF information.

[0022] As a further improvement, in step S4, the receiving end calculates the confused steering matrix V obf based on the NDP received in step S3. obf The steering matrix V obf confuses the time-space state information, specifically:

[0023] The unconfused NDP contains the long training field LTF information X required for channel probing, and the confused NDP carries QX. The process of calculating the confused steering matrix V obf after channel probing is as follows:

[0024] First, the receiving end calculates the confused channel state information CSI using the confused LTF signal QX, which is represented as:

[0025] H obf =HQ

[0026] Where H obfis the obfuscated channel state information (CSI), H is the unobfuscated channel state information, and Q is the obfuscation matrix.

[0027]

[0028] where U is the unitary matrix obtained by SVD decomposition, and V obf is the obfuscated steering matrix.

[0029] As a further improvement, in steps S4 and S5, the receiving end receives the null data packet (NDP), calculates the obfuscated steering matrix V obf , and sends it back to the router end in the form of a compressed beamforming feedback packet (CBF), without any modification to the hardware and processing procedure of the receiving end, with extremely low hardware modification cost, specifically:

[0030] In the entire protection mechanism, the router end uses the obfuscation matrix Q to obfuscate the channel probe signal, and the processing procedure of the receiving end is not modified. The receiving end normally performs channel probing and sends the calculated steering matrix V obf to the router end in the form of a compressed beamforming feedback packet (CBF), and V obf is restored at the router end, ensuring minimal modification of the receiving end and reducing cost.

[0031] As a further improvement, in step S7, the router end uses the space-time obfuscation matrix Q obtained in step S1 to restore the obfuscated steering matrix V obf obtained in step S6, to obtain the restored steering matrix V, which can perform normal beamforming without affecting normal communication, specifically:

[0032] The relationship between the obfuscated steering matrix and the unobfuscated steering matrix can be represented as:

[0033]

[0034] where V obf is the obfuscated steering matrix, is the conjugate matrix of the obfuscation matrix, and V is the unobfuscated steering matrix. Since Q is a unitary matrix, V = QV obf can be calculated to obtain the restored steering matrix V, and using the calculated restored steering matrix V to perform spatial mapping on the transmitting antenna can achieve a normal beamforming process.

[0035] The beneficial effects of the present application are as follows:

[0036] Beamforming technology is widely used in existing WiFi devices, but there is a privacy leakage problem in the beamforming process. The existing solutions have the problems of large calculation overhead, affecting communication performance, and large hardware changes, and the cost of wide application is too high. In order to solve the problems of large calculation overhead and affecting communication performance of the existing solutions, the present application utilizes the mathematical characteristics in the beamforming channel detection process, and the router end generates a space-time confusion matrix Q according to the space mapping mechanism, and only needs to multiply the space-time confusion matrix Q to restore the steering matrix after confusion, the calculation efficiency is high, the overhead is small, and the restored steering matrix can be normally used for beamforming, realizing efficient protection of personal location privacy without affecting communication performance; In order to solve the problems of large hardware changes and high cost of wide application of the existing solutions, the present application reuses the space mapping mechanism of the router end, uses the space-time confusion matrix Q as the space mapping matrix during the router beamforming channel detection, and the confusion of the signal and the restoration of the steering matrix are completed at the router end, realizing zero change of the receiving end and greatly reducing the hardware change. Compared with the prior art, the present application significantly reduces the cost of wide application, and has little effect on communication, and is suitable for actual scenes. BRIEF DESCRIPTION OF DRAWINGS

[0037] Figure 1 is a flowchart of the present application;

[0038] Figure 2 is an implementation schematic diagram of the present application. DETAILED DESCRIPTION

[0039] The technical solutions of the present application will be further described below by combining with the drawings in the specification and through specific embodiments:

[0040] The present application aims at the privacy leakage problem in the beamforming process, and aims to solve the problems of large calculation overhead, affecting communication performance and large hardware changes of the existing solutions, and high cost of wide application, and proposes a method for preventing privacy leakage in the beamforming process based on the space mapping mechanism of the router end, Figure 1 is a flowchart of the present application;

[0041] The specific implementation method of the present application is as follows:

[0042] S1, the router end obtains the bandwidth of the current network channel and the WiFi protocol, determines the subcarrier number k of the channel according to the correspondence between the bandwidth and the WiFi protocol and the subchannel number, randomly generates a unitary matrix with the size of k x k as a confusion matrix Q using QR decomposition technology, and the unitary matrix Q is used to confuse the signal used in subsequent channel detection, a new confusion matrix is generated in each beamforming process to prevent the confusion matrix from being cracked;

[0043] S2, the confusion matrix Q described in step S1 is used as a spatial mapping matrix for mapping the router transmitting antenna to the spatial stream, replacing the default direct mapping spatial mapping matrix. The signal transmitted in the channel probe in the beamforming process is the long training field LTF in the null data packet NDP, which is a fixed signal in the protocol, denoted as X. The default direct mapping spatial mapping matrix will send the LTF to the receiving end without any change. After spatial mapping using the confusion matrix Q, the LTF transmitted by the router end is a confused signal, denoted as X obf , the specific formula is:

[0044] X obf = QX

[0045] wherein X obf is the signal obtained after confusion, Q is the confusion matrix, and X is the long training field LTF.

[0046] S3, the receiving end performs the default channel probe process after receiving the confused LTF signal X obf described in step S2. After the receiving end performs the channel probe, the confused channel state information CSI is obtained, denoted as H obf . Then, the confused channel state information H obf is decomposed by SVD to obtain the confused steering matrix V obf , which is represented as:

[0047]

[0048] wherein U is the unitary matrix obtained after SVD decomposition, V obf is the confused steering matrix, and Σ is the singular value matrix denotes conjugate.

[0049] S4, according to the WiFi protocol, the bandwidth and the configuration of multiple-input and multiple-output MIMO, the receiving end selects appropriate bit number to quantize and compress the confused steering matrix V obf . The bit number for quantization has two combinations, which are 4 bits, 6 bits and 7 bits, 9 bits. Two kinds of angle information are generated and written into the beamforming feedback packet CBF, and the beamforming feedback packet CBF is sent back to the router.

[0050] S5, the router receives the beamforming feedback packet, extracts the angle information therefrom, and performs inverse quantization according to the quantization algorithm to obtain the confused steering matrix V obf .

[0051] S6, the router end uses the confusion matrix Q described in step S1 to restore the confused steering matrix. The relationship between the confused steering matrix and the steering matrix without confusion can be represented as:

[0052]

[0053] Where V obf It is the confused turning matrix. Q is the conjugate matrix of the confusion matrix, and V is the unconfined turning matrix. Since Q is a unitary matrix, we use V = QV. obf The restored steering matrix V can be calculated;

[0054] S7. The router uses the restored steering matrix V as a spatial mapping mechanism to map the transmit nature to the spatial flow, and performs the subsequent normal beamforming process.

[0055] The method for preventing privacy leaks during beamforming using the router-side spatial mapping mechanism of this invention is experimentally verified as follows:

[0056] In laboratory settings, such as Figure 2 As shown, an experimental scenario was constructed using one router and two receivers. The user moved around the router, including three states: walking, stationary, and not in the room. An attacker outside the room sniffed and parsed beamforming feedback packets (CBFs) to analyze the user's spatial information. The user performed each of the three states—walking, stationary, and not in the room—for 3 minutes each. The receivers connected to the router's WiFi and maintained a high network load by downloading large files. The attacker outside the room sniffed CBFs during beamforming and grouped the data into 50-second windows. The router initially performed beamforming normally without any defense measures, and then employed the present invention for defense, collecting 3 minutes of data for each of the three states, for a total of 9 minutes of data.

[0057] Without using the method of this invention to prevent privacy leakage during beamforming, attackers can analyze data in beamforming feedback packets and achieve an accuracy rate of up to 82.7% in determining if a user is in a room when the user is active, and up to 96.7% when the user is not in a room. However, with the method of this invention, because the data in the beamforming feedback packets sniffed by attackers is obfuscated, their accuracy rate in determining the two cases is only about 50%, which is equivalent to random judgment. Furthermore, since the network speed is evaluated at the receiving end, the defense method of this invention has almost no impact on network speed. This invention ensures that privacy is not leaked during beamforming, reduces computational overhead, minimizes the impact on communication, and requires minimal hardware modifications, making it suitable for practical application deployments.

[0058] The above is not a limitation of the present application, it should be pointed out that for those skilled in the art, without departing from the essential scope of the present application, several changes, modifications, additions or substitutions can be made, these improvements and refinements should be considered as the protection scope of the present application.

Claims

1. A method for preventing privacy leakage in a beamforming process based on a router-end space mapping mechanism, characterized in that: The method comprises the following steps: S1, generating a space-time confusion matrix Q according to a space mapping mechanism at a router end; S2, using the space-time confusion matrix Q obtained in step S1 as a space mapping matrix during channel sounding of the router beamforming; S3, performing space mapping at the router end and sending a null data packet (NDP) to a receiving end; S4, the receiving end calculates the obfuscated steering matrix V based on the NDP received in step S3 obf ; S5, the receiving end obtains the steering matrix V in step S4 obf Quantized compression is sent back to the router end in the form of compressed beamforming feedback packet CBF; S6. The router decompresses the compressed beamforming feedback packet (CBF) received in step S5 to obtain the confused steering matrix V. obf ; S7. The router end uses the space-time confusion matrix Q obtained in step S1 to the confusion turned matrix V obtained in step S6 obf reduction, to obtain the reduced turned matrix V; S8, using the restored steering matrix V for beamforming.

2. The method for preventing privacy leakage in beamforming according to claim 1, wherein, In step S1, the router end generates a space-time confusion matrix Q according to a space mapping mechanism, which is used to confuse signals sent during channel sounding, and specifically: obtaining a channel bandwidth and a protocol type; determining the number of subcarriers k of the channel according to the channel bandwidth and the protocol type; and randomly generating a unitary matrix with a size of k x k using QR decomposition.

3. The method for preventing privacy leakage in beamforming according to claim 1, wherein, In step S2, the space-time confusion matrix Q obtained in step S1 is used as a space mapping matrix during channel sounding of the router beamforming, which is used to confuse signals sent during channel sounding, and specifically: replacing a default direct mapping space mapping matrix of a router end transmitting antenna mapping to a space flow, and using the confusion matrix Q obtained in step S1 to perform space mapping on the transmitting antenna.

4. The method for preventing privacy leakage in beamforming according to claim 1, wherein, In step S4, the receiving end calculates the obfuscated steering matrix V based on the NDP received in step S3 obf Specifically, The unobfuscated NDP contains long training field (LTF) information X required for channel sounding, and the obfuscated NDP carries QX. The obfuscated steering matrix V calculated after channel sounding obf The relationship between the unobfuscated steering matrix and the obfuscated steering matrix is: V = QV obf The steering matrix V obf obfuscates the time-space state information and hides the privacy information carried in V.

5. The method for preventing privacy leakage in beamforming according to claim 1, wherein, In step S5, the receiving end transmits the turning matrix V obtained in step S4. obf After quantization and compression, the data is sent back to the router in the form of a Compressed Beamforming Feedback (CBF) packet, specifically: In the whole protection mechanism, the router end uses the confusion matrix Q to confuse the channel sounding signal, and the processing flow of the receiving end is not changed. The receiving end normally performs channel sounding and sends the calculated steering matrix V obf to the router end in the form of a compressed beamforming feedback packet CBF. obf The restoration of V 6. The method for preventing privacy leakage in beamforming according to claim 1, wherein, In step S7, the router end uses the space-time confusion matrix Q obtained in step S1 to perform confusion on the turning matrix V obtained in step S6 obf The reduction is performed to obtain the reduced turning matrix V, and the reduction is specifically as follows: Using the space-time confusion matrix Q obtained in step S1, using V = QV obf The reduced steering matrix V is calculated, and the transmit antennas are spatially mapped using the calculated reduced steering matrix V to achieve a normal beamforming process.

Images