An Adaptive Detection Method for Weak Signals Based on Stochastic Resonance
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- ZHEJIANG UNIV
- Filing Date
- 2024-07-18
- Publication Date
- 2026-06-30
Smart Images
Figure CN118945735B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to a weak signal detection method based on random resonance. Addressing the issue that when low-power wide-area networks (LPWANs) operating in similar frequency bands to Wi-Fi devices operate simultaneously at close range, weak wireless signal data packets are subject to interference from Wi-Fi. This invention proposes a weak signal detection method based on random resonance to enable coexistence between various weak wireless signals and Wi-Fi. By adding appropriate white noise, the random resonance phenomenon is utilized to enhance weak signals with specific frequencies. Furthermore, an adaptive method for adjusting random resonance parameters is proposed to modify system parameters, detect the type of weak signal exhibiting random resonance, and reserve spectrum for transmission based on the detected signal's spectral occupancy, thereby achieving coexistence between weak wireless signals and Wi-Fi. Technical Background
[0002] In the wave of modern smart homes and Internet of Things (IoT) devices, wireless communication technologies such as BLE (Bluetooth Low Energy), LoRa (Long Range Wireless Communication), and ZigBee (Short Range Wireless Communication) are becoming increasingly popular. These technologies have gained widespread application by providing low-power, high-efficiency, and reliable wireless connectivity, enabling various devices to easily interconnect and communicate. However, the potential interference and conflicts between LoRa, ZigBee, and BLE and Wi-Fi remain a significant concern. This interference can affect communication quality, data transmission rates, and even connection stability, thereby impacting the performance of IoT devices and the user experience.
[0003] To achieve coexistence between wireless signals and Wi-Fi, Wi-Fi devices first need to be able to detect the presence of devices with weak signals and stop transmitting during weak signal transmissions to reserve a channel for them, thus achieving coexistence. However, existing low-power transmission technologies transmit low energy, making them difficult for high-power devices to detect. To address this issue, this invention employs the principle of random resonance to enhance wireless signals.
[0004] Stochastic resonance is a physical phenomenon that, under certain conditions, noise of a suitable magnitude can cause an unstable state in a system, resulting in signal transitions. This transitions not only do not interfere with the original signal, but can actually amplify it, thereby enabling the system to better detect and transmit weak signals.
[0005] There are generally two ways in which stochastic resonance occurs. One is to adjust the system noise to a suitable level, but in practical applications, it is difficult to control the noise level, so this method is not considered. The other method is to adjust the system parameters. The entire environment can be considered as a bistable stochastic resonance system model. A bistable system is a dynamic system with two stable states and two different equilibrium points. The system can switch between these two states, and there is a potential barrier between them. When the initial signal and noise meet the matching conditions, the signal can successfully overcome the barrier and transition between the two potential wells, thus causing stochastic resonance. The magnitude of the potential barrier is controlled by system parameters a and b, and the occurrence of stochastic resonance can be controlled by adjusting the magnitude of these parameters.
[0006] However, suitable system parameters a and b are not easily obtained in actual use. Most existing works perform brute-force search on all possible system parameters to obtain the optimal system parameters, resulting in high time costs. Summary of the Invention
[0007] To address the problem of low signal strength during wireless signal transmission, this invention proposes a weak signal detection method based on random resonance.
[0008] This invention utilizes the principle of random resonance to enhance the signal, thereby facilitating subsequent implementation and co-transmission with Wi-Fi. The specific steps include:
[0009] S1: The raw signal acquisition unit acquires the raw signal and sends the acquired signal to the preprocessor.
[0010] S2: The preprocessor uses different preprocessing methods for different weak wireless signals in turn, and then sends the processed information to the stochastic resonance adaptor.
[0011] S3: The stochastic resonance adaptor alternately uses stochastic resonance adaptive enhancement methods for different weak wireless signals and determines the signal type and channel based on the results.
[0012] S4: A high-power Wi-Fi signal generator controls the Wi-Fi signal to yield the spectrum corresponding to the weak wireless signal being transmitted, enabling the co-transmission of various weak signals and Wi-Fi signals.
[0013] Preferably, in step S1, the original signal collected by the signal collector is mainly composed of Wi-Fi signals, and also contains LoRa, ZigBee, and BLE signals.
[0014] Preferably, in step S2, the preprocessor processes the three types of weak wireless signals as follows:
[0015] For time-varying frequency LoRa signals, the preprocessing method is to convert the time-varying LoRa signal into a fixed frequency signal, and then pass it into a low-pass filter belonging to the channel to filter out high-frequency signals before sending it into a stochastic resonant adaptive.
[0016] For BLE signals, the high-frequency signals are filtered out by a low-pass filter before being sent to a stochastic resonant adaptive oscillator.
[0017] The ZigBee signal is fed directly into the stochastic resonant adaptive amplifier without any processing.
[0018] Preferably, the step of converting the time-varying LoRa signal to a fixed frequency in S2 is as follows:
[0019]
[0020] Where θ is the frequency used by the Chirp in the LoRa signal, and the Chirp is the basic communication unit of LoRa, which is a signal whose frequency increases linearly with time. It is used to convert LoRa signals into time-varying signals of a fixed frequency. Its value is the intermediate frequency of each chirp interval in LoRa. Multiplying it with the LoRa signal cancels out the time-varying part, thereby converting LoRa signals of different frequencies into the enhanced frequency f of random resonance. g .
[0021] Preferably, the specific method for the random resonance of S2 is as follows:
[0022]
[0023]
[0024] U(x) is the potential function of the bistable system at this time, which describes the overdamped motion of the particle in the double potential well under the action of periodic driving force and noise. a and b are system parameters, both of which are constants greater than 0. x(t) is the system output signal, s(t)=Acos(2πf0t) is the weak signal to be detected, which serves as the periodic driving force of the system, and τ(t) is the noise of the system, which is Gaussian white noise.
[0025] Differentiating the potential function yields the maximum and minimum values of the curve, as well as the height difference of the potential barrier. By modifying system parameters a and b, the transition of particles in the potential well can be controlled, resulting in stochastic resonance.
[0026] Preferably, the method for detecting the type and channel of different weak wireless signals through random resonance in S3 is as follows:
[0027] The enhancement method for LoRa signals is to iterate the enhancement frequency f of each channel. gThe corresponding random resonance parameters a and b are used to perform random resonance, and the existence of LoRa signal in the corresponding channel is determined based on whether random resonance occurs.
[0028] For ZigBee and BLE signals, the adaptive regulator built into the signal enhancer is used to transform the acquisition of the optimal system parameters for random resonance into a constrained nonlinear optimization problem to find the corresponding signal type and channel.
[0029] Preferably, the steps in step S3 for solving the constrained nonlinear optimization problem of finding the optimal system parameters for the signal are as follows:
[0030] minf(b) (4)
[0031] c(b) = L b -b≤0 (5)
[0032] minφ k (b)=f(b)+σ k ∑g(c(b)) (6)
[0033] g(c(b)) = max(0, c(b)) 2 (7)
[0034] Wherein, the constraint problem minf(b) represents the negative value of the amplitude of the pixel set in the frequency band amplified by random resonance in the CWT spectrum obtained using wavelet transform, and the lower bound L in the constraint condition is... b Let a certain value of b during the rising phase be used as a constraint condition. In the above equation, this is called the external penalty function f(b), σ. k This is called the penalty factor, and in each iteration we increase σ. k Then solve the unconstrained problem. The results of each iteration are combined into a sequence, and the limit of this sequence is the solution to the original constrained problem.
[0035] Preferably, the high-power Wi-Fi signal generator in S4 controls the Wi-Fi signal to yield the spectrum of the weak wireless signal being transmitted in the following way:
[0036] A high-power Wi-Fi signal generator emits a high-power Wi-Fi signal, forcing nearby Wi-Fi signals at the same frequency to back off, thus reserving spectrum for weaker signals.
[0037] This invention addresses the problem of low signal strength during wireless signal transmission by proposing a weak signal detection method based on stochastic resonance. The method enhances the signal strength using the principle of stochastic resonance, dynamically adjusts system parameters by using the output signal of a bistable system as a feedback signal, and employs a penalty function method to enable the stochastic resonance model to quickly converge to the optimal system parameters, avoiding tedious brute-force searches and facilitating subsequent implementation for co-transmission with Wi-Fi.
[0038] The advantages of this invention are: it can perform random resonance detection on various weak wireless signals even without knowing the type of the transmitted weak wireless signal, achieving automated detection; it is easy to use; it uses the penalty function method to adaptively adjust system parameters based on the results of random resonance; it is applicable to detecting weak wireless signals in different ranges, has a wide range of applications, and a fast response. Each module performs its own function, the structure is clear, and it is easy to use. Attached Figure Description
[0039] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0040] Figure 1 This is a system diagram for implementing the method of the present invention. Detailed Implementation
[0041] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0042] Example 1
[0043] Reference Figure 1 To better identify adversarial examples, this invention provides a weak signal detection method based on stochastic resonance, specifically including the following steps:
[0044] S1: The raw signal acquisition unit acquires the raw signal and sends the acquired signal to the preprocessor.
[0045] The original signal collected by the signal collector in step S1 is mainly composed of Wi-Fi signals, and also contains LoRa, ZigBee, and BLE signals.
[0046] Reference Figure 1S2: The preprocessor uses different preprocessing methods for different weak wireless signals in turn, and then the processed information is sent to the stochastic resonance adaptor.
[0047] The preprocessor in step S2 handles the three types of weak wireless signals as follows:
[0048] For time-varying frequency LoRa signals, the preprocessing method involves converting the time-varying LoRa signal into a fixed-frequency signal using a trigonometric transform. The steps are as follows:
[0049]
[0050] Where θ is the frequency used by the Chirp in the LoRa signal, and the Chirp is the basic communication unit of LoRa, which is a signal whose frequency increases linearly with time. It is used to convert LoRa signals into time-varying signals of a fixed frequency. Its value is the intermediate frequency of each chirp interval in LoRa. Multiplying it with the LoRa signal cancels out the time-varying part, thereby converting LoRa signals of different frequencies into the enhanced frequency f of random resonance. g .
[0051] After preprocessing, the LoRa signal is fed into a low-pass filter belonging to the channel, and high-frequency signals are filtered out before being sent to a stochastic resonant adaptive oscillator.
[0052] For BLE signals, the high-frequency signals are filtered out by a low-pass filter before being sent to a stochastic resonant adaptive oscillator.
[0053] The ZigBee signal is fed directly into the stochastic resonant adaptive amplifier without any processing.
[0054] Reference Figure 1 S3: The stochastic resonance adaptor alternately uses stochastic resonance adaptive enhancement methods for different weak wireless signals, and determines the signal type and channel based on the results.
[0055] The specific method of stochastic resonance is as follows:
[0056]
[0057]
[0058] U(x) is the potential function of the bistable system at this time, which describes the overdamped motion of the particle in the double potential well under the action of periodic driving force and noise. a and b are system parameters, both of which are constants greater than 0. x(t) is the system output signal, s(t)=Acos(2πf0t) is the weak signal to be detected, which serves as the periodic driving force of the system, and τ(t) is the noise of the system, which is Gaussian white noise.
[0059] Differentiating the potential function yields the maximum and minimum values of the curve, as well as the height difference of the potential barrier. By modifying system parameters a and b, the transition of particles in the potential well can be controlled, resulting in stochastic resonance.
[0060] The enhancement method for LoRa signals is to iterate the enhancement frequency f of each channel. g The corresponding random resonance parameters a and b are used to perform random resonance, and the existence of LoRa signal in the corresponding channel is determined based on whether random resonance occurs.
[0061] For ZigBee and BLE signals, the adaptive regulator built into the signal enhancer is used to transform the acquisition of the optimal system parameters for random resonance into a constrained nonlinear optimization problem to find the corresponding signal type and channel.
[0062] The steps for solving the constrained nonlinear optimization problem of finding the optimal system parameters for the signal are as follows:
[0063] minf(b) (4)
[0064] c(b) = L b -b≤0 (5)
[0065] minφ k (b)=f(b)+σ k ∑g(c(b)) (6)
[0066] g(c(b)) = max(0, c(b)) 2 (7)
[0067] Wherein, the constraint problem minf(b) represents the negative value of the amplitude of the pixel set in the frequency band amplified by random resonance in the CWT spectrum obtained using wavelet transform, and the lower bound L in the constraint condition is... b Let a certain value of b during the rising phase be used as a constraint condition. In the above equation, this is called the external penalty function f(b), σ. k This is called the penalty factor, and in each iteration we increase σ. k Then solve the unconstrained problem. The results of each iteration are combined into a sequence, and the limit of this sequence is the solution to the original constrained problem.
[0068] Reference Figure 1 S4; A high-power Wi-Fi signal generator emits a high-power Wi-Fi signal, forcing nearby Wi-Fi signals of the same frequency to retreat, thereby reserving spectrum for weak signals and enabling the co-transmission of various weak signals and Wi-Fi signals.
[0069] Table 1
[0070] SNR (dB) -25 -15 -10 -5 0 Raw 0 0 0 0 0.08 EmBee 0.2 0.3 0.4 0.4 0.5 Stochastic Resonance 0.94 0.95 0.97 0.98 1
[0071] Table 1 shows the detection results of weak wireless signals under different signal-to-noise ratios based on different methods. It can be seen that the wireless signal detection method based on random resonance can effectively detect adversarial examples.
[0072] Table 2 shows the Wi-Fi throughput loss caused by this invention.
[0073] Data Rate (Mbps) 50 43 33 24 12 Original 24.44 22.35 16.78 15.22 13.22 Stochastic Resonance 22.33 20.12 14.58 13.84 12.23 Throughout loss 15.54 9.96 13.12 8.36 7.66
[0074] As can be seen from the table showing the Wi-Fi throughput loss based on random resonance, the present invention results in less loss to Wi-Fi.
[0075] Example 2
[0076] Reference Figure 1 This embodiment relates to a weak signal detection system based on stochastic resonance, comprising: a raw signal sampler, a preprocessor, a stochastic resonance enhancement module, and a high-power Wi-Fi signal generator. The raw signal sampler is responsible for acquiring the initial signal. The preprocessor is responsible for processing the raw signal, using different preprocessing methods for LoRa, ZigBee, and BLE signals, and applying them iteratively. The stochastic resonance enhancement module is responsible for performing stochastic resonance enhancement on the preprocessed signal, using different methods for different signals. Subsequently, the Wi-Fi signal generator emits a high-power Wi-Fi signal after detecting the signal, thus clearing the path for the weak signal. This invention addresses the issue that when LoRa, ZigBee, and BLE wireless communication technologies operating in similar frequency bands to Wi-Fi operate simultaneously with Wi-Fi devices at close range, their data packets are subject to Wi-Fi interference. By using the output signal of the bistable system as a feedback signal, the system parameters are dynamically adjusted, allowing the stochastic resonance model to quickly converge to the optimal system parameters, avoiding tedious brute-force searches. Then, based on the detected spectrum occupancy of weak signals, spectrum is reserved for the transmission of weak signals, thus enabling coexistence with Wi-Fi.
[0077] The embodiments described in this specification are merely examples of implementations of the inventive concept. The scope of protection of this invention should not be considered as limited to the specific forms stated. The scope of protection of this invention also extends to equivalent technical means that can be conceived by those skilled in the art based on the inventive concept.
Claims
1. A method for detecting weak signals based on stochastic resonance, characterized in that, Includes the following steps: S1: The raw signal acquisition unit acquires the raw signal and sends the acquired signal to the preprocessor; S2: The preprocessor uses different preprocessing methods for different weak wireless signals in turn, and then the processed signals are sent to the random resonance adaptor. S3: The stochastic resonance adaptor alternately uses stochastic resonance adaptive enhancement methods for different weak wireless signals and determines the signal type and channel based on the results; The specific method for the aforementioned stochastic resonance is as follows: (2) (3) For the potential function of the bistable system, the overdamped motion of a particle in a double-well potential under the action of periodic driving force and noise is described, a and b are system parameters, both are constants greater than 0; For the system output signal, For the weak signal to be detected, as the periodic driving force of the system, is the noise of the system, which is a Gaussian white noise; The derivative of the potential function yields the maximum and minimum values of the curve, as well as the height difference of the potential barrier. By modifying the system parameters a and b, the transition of particles in the potential well can be controlled, resulting in random resonance. The methods for detecting different types and channels of weak wireless signals using random resonance are as follows: The enhancement mode for the LoRa signal is to iterate the enhancement frequency of each channel The corresponding random resonance parameter to perform random resonance and determine whether the LoRa signal of the corresponding channel exists according to whether random resonance occurs For ZigBee and BLE signals, the adaptive regulator built into the signal enhancer is used to transform the acquisition of the optimal system parameters for random resonance into a constrained nonlinear optimization problem in order to find the corresponding signal type and channel. The steps for solving the constrained nonlinear optimization problem of finding the optimal system parameters for the signal are as follows: (4) (5) (6) (7) Among them, the constraint problem The lower bound in the constraint is the negative value of the amplitude of the pixel set in the CWT spectrum obtained by wavelet transform, which is amplified through the random resonance. Set as a certain period of growth The value is used as a constraint condition in the above equation. Called External penalty function, This is called the penalty factor, and it increases in each iteration. Then solve formula (6); the results of each iteration will be combined into a sequence, and the limit of this sequence is the solution to the original constraint problem; S4: A high-power Wi-Fi signal generator controls the Wi-Fi signal to yield the spectrum corresponding to the weak wireless signal being transmitted, enabling the co-transmission of various weak signals and Wi-Fi signals.
2. The weak signal detection method based on stochastic resonance according to claim 1, characterized in that: The original signal collected by the signal collector in step S1 is mainly composed of Wi-Fi signals, and also contains LoRa, ZigBee, and BLE signals.
3. The weak signal detection method based on stochastic resonance according to claim 1, characterized in that: The preprocessor in step S2 handles the three types of weak wireless signals as follows: For time-varying frequency LoRa signals, the preprocessing method is to convert the time-varying LoRa signal into a fixed frequency signal, and then pass it into a low-pass filter belonging to the channel to filter out high-frequency signals before sending it into a stochastic resonant adaptive. For BLE signals, the high-frequency signals are filtered out by a low-pass filter before being sent to a stochastic resonant adaptive oscillator. The ZigBee signal is fed directly into the stochastic resonant adaptive amplifier without any processing.
4. The weak signal detection method based on stochastic resonance according to claim 3, characterized in that: The step of converting the time-varying LoRa signal to a fixed frequency in step S2 is as follows: (1) in This refers to the frequency used by the Chirp in the LoRa signal. The Chirp is the basic communication unit of LoRa, which is a signal whose frequency increases linearly with time. It is used to convert LoRa signals into time-varying signals of a fixed frequency. Its value is the intermediate frequency of each chirp interval in LoRa. Multiplying it with the LoRa signal cancels out the time-varying part, thereby converting LoRa signals of different frequencies into the enhanced frequency of random resonance. .
5. The weak signal detection method based on stochastic resonance according to claim 1, characterized in that: In step S4, the high-power Wi-Fi signal generator controls the Wi-Fi signal to yield the spectrum of the weak wireless signal being transmitted in the following way: A high-power Wi-Fi signal generator emits a high-power Wi-Fi signal, forcing nearby Wi-Fi signals at the same frequency to back off, thus reserving spectrum for weaker signals.