Intelligent sensor based on physical unclonable function

Abstract

The invention belongs to the technical field of sensors, and particularly relates to an intelligent sensor based on a physical unclonable function, which dynamically identifies the situation type of the environment where a current PUF circuit is located by calculating the change rate and the change acceleration of a plurality of environmental physical parameters of the PUF circuit in a preset time window, and responds to a strong mutation situation signal, so that the current PUF circuit can be identified. Calling a specific compensation vector associated with the current environment strong abrupt change situation in a pre-stored compensation vector set, applying the specific compensation vector to a response signal output end of the PUF circuit, monitoring and verifying the stability of an output response signal, if the stability reaches the standard, associatively storing the current environment strong abrupt change situation and the specific compensation vector, and if the stability reaches the standard, storing the current environment strong abrupt change situation and the specific compensation vector. Otherwise, the specific compensation vector is revoked, and a standard correction process is started, so that an intelligent sensor architecture which actively senses an environment sudden change trend, implements accurate pre-correction in advance and has a self-optimization capability is created, and high-stability and high-reliability authentication of the PUF identity identifier in a full life cycle is realized.

Description

Technical Field

[0001] This invention belongs to the field of sensor technology, specifically a smart sensor based on physically unclonable functions. Background Technology

[0002] With the widespread adoption of IoT technology, smart sensors, as core components connecting the physical world and digital systems, face critical challenges in terms of security authentication and data trustworthiness. Traditional solutions relying on software keys or external security chips are vulnerable to physical attacks, key leaks, and cloning, making it difficult to achieve hardware-level intrinsic security.

[0003] Physically Unclonable Function (PUF) technology utilizes the inherent microscopic differences in the semiconductor manufacturing process to generate a unique and repeatable response signal for each device, forming an uncopyable hardware fingerprint.

[0004] To provide an effective path for native trusted authentication of device identity, existing PUF authentication schemes typically collect initial responses as reference features in a stable environment, and then compare and correct real-time responses using error correction codes and other technologies in subsequent authentication to address response drift issues caused by environmental fluctuations such as temperature and voltage.

[0005] However, such methods are inherently lagging correction mechanisms, initiating the compensation process only after a response deviation is detected. This leads to initial authentication failures in practical applications when the sensor encounters sudden environmental changes due to significant drift in the PUF response within a short period. Furthermore, frequent or inappropriate correction operations incur unnecessary computational overhead and may even introduce additional errors due to overcorrection.

[0006] Ultimately, existing technologies lack the ability to predict and classify dynamic environmental changes, making it difficult to balance real-time certification, system energy efficiency, and long-term stability, which constitutes the main bottleneck for large-scale applications. Summary of the Invention

[0007] To overcome the shortcomings of the prior art, embodiments of the present invention provide an intelligent sensor based on physically unclonable functionality, which can effectively solve the problems involved in the prior art.

[0008] The objective of this invention can be achieved through the following technical solution: an intelligent sensor based on physically unclonable functionality, comprising: an environmental parameter monitoring module, an environmental situation awareness module, a compensation vector loading module, and a compensation effect verification module.

[0009] The environmental parameter monitoring module is connected to the environmental situation awareness module, the environmental situation awareness module is connected to the compensation vector loading module, and the compensation vector loading module is connected to the compensation effect verification module.

[0010] The environmental parameter monitoring module monitors multiple environmental physical parameters of the PUF circuit and calculates the rate of change and acceleration of change of each environmental physical parameter within a predetermined time window.

[0011] The environmental situation awareness module dynamically identifies the situation type of the current environment in which the PUF circuit is located based on the rate of change and the acceleration of change. The situation type is one of the following: normal situation, non-stationary situation, and strong sudden change situation.

[0012] The compensation vector loading module, in response to a strong mutation situation signal, calls a specific compensation vector from the pre-stored compensation vector set that is associated with the current strong mutation situation in the environment, and applies the specific compensation vector to the response signal output terminal of the PUF circuit.

[0013] The compensation effect verification module monitors the stability of the output response signal over multiple consecutive sampling periods after applying a specific compensation vector. If the stability meets the standard, it associates and stores the current strong change situation in the environment with the specific compensation vector; otherwise, it cancels the specific compensation vector and enables the standard correction process.

[0014] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects: (1) The present invention monitors the rate of change and acceleration of change of environmental parameters in real time, and based on the hierarchical situation recognition mechanism, actively loads a specific compensation vector when a strong sudden change situation is identified. This pre-warning and active compensation mechanism can intervene before or in the early stage of the PUF response drifts significantly, effectively overcomes the lag of the existing post-correction, and significantly improves the real-time success rate of identity authentication and the reliability of the response signal.

[0015] (2) This invention precisely divides the environmental situation into three types: normal, non-stationary and strong mutation, and sets different processing strategies. It only enables the loading of compensation vectors when a strong mutation situation is detected, avoiding the need to start a high-overhead correction process for gradual or short-term fluctuations. This significantly reduces the computational load and resource consumption of the system while ensuring the robustness of the certification.

[0016] (3) Through the compensation effect verification and system self-optimization mechanism, this invention can continuously learn and incrementally update the compensation vector set and decision parameters based on the compensation effect records in the actual environment. It can dynamically adapt to new or extreme environmental conditions and continuously optimize the compensation strategy, thereby significantly enhancing the long-term certification robustness and environmental adaptability of the smart sensor under complex and variable working conditions. Attached Figure Description

[0017] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of the module connection of the present invention.

[0019] Figure 2 This is a logical diagram illustrating the dynamic identification of environmental situation types in this invention.

[0020] Figure 3 This is a schematic diagram illustrating the self-optimization logic of the system of the present invention. Detailed Implementation

[0021] 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.

[0022] Reference Figure 1 As shown, the present invention provides an intelligent sensor based on physically unclonable functionality, comprising: an environmental parameter monitoring module, an environmental situation awareness module, a compensation vector loading module, and a compensation effect verification module.

[0023] The environmental parameter monitoring module is connected to the environmental situation awareness module, the environmental situation awareness module is connected to the compensation vector loading module, and the compensation vector loading module is connected to the compensation effect verification module.

[0024] The environmental parameter monitoring module monitors multiple environmental physical parameters of the PUF circuit and calculates the rate of change and acceleration of change of each environmental physical parameter within a predetermined time window.

[0025] In one embodiment of the present invention, the monitoring of multiple environmental physical parameters of the PUF circuit is specifically achieved by integrating multiple environmental monitoring elements in the vicinity of the PUF circuit, so as to ensure that the sensed environmental parameters can reflect the actual microenvironmental conditions of the PUF circuit to the greatest extent.

[0026] The multiple monitoring elements correspond to the environmental physical parameters of interest, including but not limited to: temperature sensors, humidity sensors, and power voltage monitoring sensors.

[0027] As an example, the "proximate area" refers to the multiple monitoring elements and the PUF circuit located on the same chip substrate or inside the same chip package, and the physical distance between them and the core area of ​​the PUF functional unit is set to a preset ratio smaller than the chip package size, so as to ensure that the monitoring data can effectively characterize the instantaneous changes in the local environment where the PUF circuit is located.

[0028] The environmental situation awareness module dynamically identifies the situation type of the current environment in which the PUF circuit is located based on the rate of change and the acceleration of change. The situation type is one of the following: normal situation, non-stationary situation, and strong sudden change situation.

[0029] Given that the rate of change of environmental physical parameters reflects the intensity of environmental disturbances, while the acceleration of change reveals the drastic nature of the disturbance: the non-stationary state is characterized by environmental physical parameters deviating from the steady state, with the rate of change exceeding the threshold, but the change process is relatively gradual, and the kinetic energy of the disturbance has not been fully released; the strong abrupt change state is characterized by environmental physical parameters exceeding the standard rate of change, and the acceleration of change further exceeding the threshold, which means that the intensity of the disturbance is rapidly amplifying, and the system is facing an instantaneous impact.

[0030] Based on this, refer to Figure 2 As shown in a preferred embodiment of the present invention, the dynamic identification process of the environmental situation type includes: setting corresponding dynamic thresholds for the rate of change and acceleration of change of each environmental physical parameter.

[0031] When the rate of change of all environmental physical parameters is less than their corresponding dynamic threshold, it is identified as a normal situation.

[0032] Conversely, if the rate of change of any environmental physical parameter is greater than or equal to its corresponding dynamic threshold, and the acceleration of change of all environmental physical parameters is less than its corresponding dynamic threshold, then it is identified as a non-stationary state.

[0033] If the acceleration of change of at least one environmental physical parameter is greater than its corresponding dynamic threshold for acceleration change, it is identified as a strong abrupt change situation.

[0034] In a preferred embodiment of the present invention, the process of setting the dynamic threshold includes: continuously storing the time series of each environmental physical parameter and its corresponding PUF response drift within a historical period.

[0035] The rate of change and acceleration sequence of each environmental physical parameter are extracted based on an adaptive sliding window.

[0036] It should be noted that the size of the sliding time window is dynamically adjusted according to the seasonal characteristics of the environmental physical parameters. A shorter window size is used in seasons with frequent parameter fluctuations, and a longer window size is used in seasons with stable parameters.

[0037] The standard deviation of the rate of change and acceleration of change of each environmental physical parameter, as well as the correlation coefficient between the change of each environmental physical parameter and the drift of the PUF response, are calculated in parallel.

[0038] As an example, the correlation coefficient between the changes in various environmental physical parameters and the drift of the PUF response is calculated using the Pearson correlation coefficient. Specifically, the changes in environmental physical parameters are recorded in unit time steps within a predetermined time window.

[0039] The PUF response drift was recorded at exactly the same time point. The offset was defined as the Hamming distance between the output response signal and the reference response signal pre-stored by the sensor, where the reference response signal is the same as the subsequent reference identity feature attribute.

[0040] Simultaneously construct the environmental physical parameter change sequence and the PUF response drift sequence.

[0041] The covariance of the two sequences is calculated to quantify the trend of the coordinated change in magnitude between environmental parameter changes and PUF response drift.

[0042] The product of the standard deviations of the environmental physical parameter change sequence and the PUF response drift sequence is used as the denominator, and the covariance of the two sequences is used as the numerator to perform a ratio calculation, thereby eliminating the influence caused by the difference in the dimensions and numerical dispersion of the two sequences.

[0043] The final result is a scalar value between -1 and +1, which is the Pearson correlation coefficient.

[0044] In other embodiments, implementers may also use the Spearman rank correlation coefficient method to capture nonlinear relationships according to the actual application requirements, or the maximum information coefficient method to fully explore the complex dependencies between parameters, which will not be elaborated here.

[0045] The normalized standard deviation is multiplied by the correlation coefficient, and then amplified by a configurable adjustment factor to generate specific dynamic thresholds for the rate of change and acceleration of change of each environmental physical parameter. The normalized standard deviation is obtained through the following process: the standard deviation of the environmental physical parameter within a predetermined time window is compared with the baseline standard deviation obtained statistically over historical periods. This ratio effectively characterizes the significance of the current short-term fluctuations relative to the long-term normal fluctuation level.

[0046] The adjustment coefficient is determined by rigorous experimental calibration of the smart sensor before it leaves the factory. The specific process is as follows: Before the sensor leaves the factory, it undergoes a test sequence containing various environmental changes in intensity in a standard test chamber.

[0047] Using a series of candidate adjustment coefficients, backtesting was performed on the above test process to obtain the false negative rate and false positive rate corresponding to each candidate adjustment coefficient. Here, the false negative rate refers to the probability that a strong mutation has occurred but the system has not triggered an early warning; the false positive rate refers to the probability that the system has falsely triggered a strong mutation early warning when the environment is only in a non-stationary state or in a normal fluctuation.

[0048] Using the condition that both the false positive and false negative rates are less than the experimentally calibrated upper limit as the core constraint, all solutions that meet the condition are selected from a series of candidate adjustment coefficients.

[0049] From the solution set that meets the conditions, the median value is selected as the initial configuration value of the adjustment coefficient, thereby establishing an optimal balance between sensitivity and stability.

[0050] It should be added that the above dynamic threshold generation logic integrates the objective fluctuations of the environment and the subjective sensitivity of the PUF circuit. Specifically, the normalized standard deviation quantifies the recent relative long-term fluctuation intensity of a certain environmental physical parameter. For environments with obvious and drastic fluctuations in the near term, the warning threshold will be automatically widened to avoid frequent false alarms due to background noise. For stable environments, the threshold will be narrowed accordingly to maintain high sensitivity to minor disturbances.

[0051] The correlation coefficient quantifies the influence of changes in a certain environmental physical parameter on PUF response drift. For parameters that are strongly correlated with PUF drift, the system will assign them a higher monitoring weight; while for parameters that are weakly correlated, the system will reduce their monitoring weight.

[0052] The adjustment coefficient represents a global trade-off between the false alarm rate and the false alarm rate of the sensor. A smaller adjustment coefficient indicates that the system adopts a more sensitive early warning strategy, tending to reduce the risk of false alarms; a larger adjustment coefficient represents a more robust strategy, aiming to suppress false alarms to improve efficiency and user experience. Furthermore, this coefficient can be adaptively fine-tuned based on the analysis of historical failure cases, thereby achieving self-optimization of system performance.

[0053] The dynamic threshold is obtained by first perceiving the environmental background through normalized standard deviation, then identifying key threats through correlation coefficients, and finally adjusting the coefficients to uniformly regulate the security strategy. The final output is no longer a fixed value, but an intelligent benchmark that dynamically changes with environmental fluctuations, PUF sensitivity, and system strategy. This fundamentally solves the technical problem that traditional fixed threshold solutions cannot simultaneously take into account adaptability, accuracy, and reliability.

[0054] The compensation vector loading module, in response to a strong mutation situation signal, calls a specific compensation vector from a pre-stored compensation vector set that is associated with the current strong mutation situation in the environment, and applies the specific compensation vector to the response signal output terminal of the PUF circuit.

[0055] In a preferred embodiment of the present invention, the process of constructing the pre-stored compensation vector set includes: applying multiple preset environmental physical parameter mutation modes sequentially in a controlled experimental environment.

[0056] In each mutation mode, the raw response signal of the PUF circuit and its corresponding environmental physical parameter change rate and change acceleration data are collected.

[0057] With the goal of achieving satisfactory response signals before and after compensation and minimizing the cumulative adjustment of compensation parameters, compensation parameters for each mutation mode are determined through iterative optimization. The specific process is as follows: Initialize the compensation parameter vector as the baseline value, and execute the following loop for each mutation mode: Apply the initial compensation parameters to the PUF response signal and obtain the compensated PUF response signal.

[0058] Authentication is performed on the compensated PUF response signal: First, the compensated PUF response signal is aligned bit-level with the pre-stored reference response signal.

[0059] Next, perform an XOR operation on each bit and count the total number of inconsistent bits.

[0060] Considering the inherent stability characteristics of the PUF circuit, it is necessary to extract the flip probability of each bit in the reference response signal during the registration phase test, arrange the flip probability sequence in descending order, assign lower weights to a preset number of bits in the flip probability sequence, assign higher weights to a preset number of bits in the sequence, assign median weights to bits in the middle of the sequence, perform weighted fusion processing on each inconsistent bit, and use the ratio of the weighted result to the total number of bits as the bit error rate.

[0061] For example, the bits in the first 30% of the flip probability sequence are assigned a weight of 0.2, the middle 40% are assigned a weight of 0.3, and the last 30% are assigned a weight of 0.5, and normalization is used to ensure that the total weight is 1.

[0062] The difference between 1 and the bit error rate is used as the confidence level for identity authentication.

[0063] If the confidence level of identity authentication reaches the high end of the preset authentication maintenance range, record the current compensation parameter and the corresponding total parameter adjustment. The total parameter adjustment is specifically the sum of the ratios of the adjusted values ​​of each compensation parameter to their original values.

[0064] If the target is not met, adjust the values ​​of each component in the compensation parameter vector according to the preset step size strategy.

[0065] Repeat the above process until the response signal meets the standard. Finally, from all the compensation parameter combinations that meet the response standard, select the parameter combination that minimizes the total amount of compensation parameter adjustment as the compensation parameter for this mutation mode.

[0066] A mapping relationship is established between environmental mutation modes and their corresponding compensation parameters, and these relationships are stored to form a set of compensation vectors.

[0067] In a preferred embodiment of the present invention, the specific compensation vector invocation process includes: constructing a first-order abrupt change situation feature matrix of the current environment, wherein each column of the matrix corresponds to an environmental physical parameter, and the column vector is composed of the rate of change of the parameter and the over-threshold value of the acceleration of change.

[0068] Each environmental mutation mode in the pre-stored compensation vector set is converted into a sample matrix.

[0069] Calculate the similarity between the first-order mutation situation matrix and each sample matrix, mark the sample matrix with the highest similarity as the candidate matrix, and use its corresponding compensation vector as the candidate compensation vector.

[0070] As an example, the similarity between the first-order mutation situation matrix and each sample matrix can be calculated using cosine similarity expansion. In other embodiments, the implementer may also use the Frobenius norm to measure matrix differences, or perform a reverse mapping through the Euclidean distance between matrices to characterize similarity.

[0071] The difference between the first-order mutation situation feature matrix and the candidate sample matrix is ​​calculated to obtain the environmental physical parameter deviation matrix.

[0072] The deviation values ​​of each parameter in the environmental physical parameter deviation matrix are sorted to identify key environmental physical parameters whose deviation exceeds a set threshold.

[0073] Based on the correlation characteristics between the key environmental physical parameters and the compensation vector, the candidate compensation vector is adjusted specifically to obtain the final specific compensation vector.

[0074] In a preferred embodiment of the present invention, the parameter-specific adjustment of the candidate compensation vector includes: based on the iterative optimization process, extracting the proportional relationship between the historical change data of the key environmental physical parameters and the corresponding change amount of the compensation parameters. The specific extraction process is as follows: based on the historical data accumulated in the iterative optimization process, extracting the pairing data of the change amount of environmental physical parameters and the corresponding change amount of the compensation parameters under each mutation mode.

[0075] For each key environmental physical parameter, perform the following analysis: calculate the absolute value sequence of the historical changes of the parameter and the absolute value sequence of the changes of the corresponding compensation parameter components.

[0076] The ratio coefficient between the two is obtained by least squares linear fitting.

[0077] The statistical significance of the proportion coefficient is verified, and the proportion coefficients that meet the significance level are retained. If there are multiple proportion coefficients that meet the requirements, the largest proportion coefficient is selected. If there are no proportion coefficients that meet the significance level, the conservative default proportion coefficient is adopted, and its value is the median of the proportion coefficients in historical successful cases.

[0078] The statistical significance verification mentioned above uses the t-test method, with a significance level of 0.05 as the judgment criterion. This is a standard setting in the field of statistics, and the specific testing process is existing technology, so it will not be described in detail here.

[0079] Based on the aforementioned proportional relationship and the actual deviation values ​​of the current key environmental physical parameters, the adjustment component of the compensation vector is calculated.

[0080] The adjustment component is superimposed with the corresponding parameter component in the candidate compensation vector to achieve directional fine-tuning of the candidate compensation vector.

[0081] This invention, through real-time monitoring of the rate and acceleration of change of environmental parameters, and based on a hierarchical situation recognition mechanism, actively loads a specific compensation vector when a strong abrupt change situation is identified. This pre-warning and proactive compensation mechanism can intervene before or in the early stages of a significant drift in the PUF response, effectively overcoming the lag of existing post-correction methods and significantly improving the real-time success rate of identity authentication and the reliability of response signals.

[0082] The compensation effect verification module monitors the stability of the output response signal within multiple consecutive sampling periods after applying a specific compensation vector. If the stability meets the standard, it associates and stores the current strong change situation in the environment with the specific compensation vector; otherwise, it cancels the specific compensation vector and enables the standard correction process.

[0083] In a preferred embodiment of the present invention, the monitoring of the stability of the output response signal includes: collecting the output response signal of the PUF circuit for each sampling period and performing identity authentication.

[0084] A time series analysis is performed on the identity authentication confidence scores of multiple consecutive sampling periods to determine whether the identity authentication confidence scores exhibit convergence characteristics within a preset authentication maintenance interval. The convergence characteristics refer to the fact that the volatility of the confidence score sequence gradually decreases and the overall trend approaches a higher confidence level within the preset authentication maintenance interval. The specific judgment process is as follows: the identity authentication confidence score sequence is divided into two adjacent time windows.

[0085] Calculate the variance of the built-in reliability for the two time windows before and after.

[0086] If the variance of the rear window is less than the variance of the front window, and the variance of the rear window does not exceed the preset quantile of the variance of the stability index in historical successful cases, then convergence is deemed to have occurred. The preset quantile can be exemplarily 75%, but implementers can customize it according to the statistical significance requirements for stability convergence in the system design. Alternatively, it can be designed as a dynamic parameter, adaptively adjusting based on the dispersion of historical successful case data. When the data distribution is concentrated, the quantile is automatically increased to improve discrimination accuracy. When the data distribution is dispersed, the quantile is appropriately decreased to ensure system availability.

[0087] If convergence characteristics are observed, the stability of the output response signal is confirmed to meet the standard.

[0088] Specifically, the introduction of convergence feature judgment means that there is no need to wait for the identity authentication confidence to reach the highest point. As long as the system is rapidly stabilizing in a positive direction, the pre-correction can be judged to be successful earlier, thereby achieving intelligent evaluation.

[0089] In a preferred embodiment of the present invention, the standard calibration process includes: collecting the raw response signal of the PUF circuit for each sampling period.

[0090] Multiple original response signals are merged to obtain a composite response signal.

[0091] As an example, the merging process can be implemented by calculating the average or weighted average of multiple original response signals. In other embodiments, the implementer can make a majority decision on multiple original response signals and determine the final bit value bit by bit. Alternatively, the original response signal with the best signal quality among multiple sampling periods can be selected, which will not be elaborated further.

[0092] Using the auxiliary data generated during the sensor registration phase and pre-stored locally, the comprehensive response signal is decoded and reconstructed to generate the reconstructed response signal to be verified.

[0093] It should be noted that the decoding and reconstruction process utilizes the error correction capability of the error correction code. Essentially, it is a process of correcting the noisy PUF response signal to the original registered state through auxiliary data. The specific execution steps are as follows: The integrated response signal and the auxiliary data read from the local secure memory are synchronously input into the decoding logic.

[0094] Using the auxiliary data as the decoding key, an error correction code decoding algorithm is executed on the integrated response signal. This algorithm automatically identifies and locates the erroneous bits in the integrated response signal by comparing the integrated response signal with the verification information in the auxiliary data.

[0095] Based on the error pattern output by the decoding algorithm, the erroneous bits identified in the synthesized response signal are flipped and corrected to obtain the reconstructed response signal.

[0096] The identity authentication is performed on the current reconstruction response signal to be verified, and the current identity authentication confidence level is obtained.

[0097] If the current confidence level of identity authentication is at the high end of the preset authentication maintenance interval, then identity authentication is completed. As a preferred option, the high end of the preset authentication maintenance interval specifically refers to the median of the interval, plus one to two times the standard deviation of the interval.

[0098] Reference Figure 3 As shown, in a preferred embodiment of the present invention, the system further includes a self-optimization process: the verification results after each compensation vector loading and execution, including environmental strong mutation situation characteristic data, the called compensation vector identifier and stability verification results, are packaged together into an authentication event record.

[0099] Based on the stability verification results, the authentication event records are classified and stored in different partitions of the historical case library. Records that meet the stability standard are stored in the successful case partition, and records that do not meet the stability standard are stored in the failed case partition.

[0100] Based on the accumulated authentication event records in the historical case library, perform at least one of the following optimization operations: statistically analyze the fluctuation characteristics of environmental parameters in the failure case partition and adjust the adjustment coefficient in the dynamic threshold calculation accordingly.

[0101] Extract newly emerging environmental situation patterns from the successful case partitions and trigger incremental updates to the pre-stored compensation vector set.

[0102] Through the compensation effect verification and system self-optimization mechanism, the embodiments of the present invention can continuously learn and incrementally update the compensation vector set and decision parameters based on the compensation effect records in the actual environment. It can dynamically adapt to new or extreme environmental conditions and continuously optimize the compensation strategy, thereby significantly enhancing the long-term certification robustness and environmental adaptability of smart sensors under complex and variable working conditions.

[0103] In a preferred embodiment of the present invention, in response to a non-stationary situation signal, rapid sampling and situation prediction processing is performed: the sampling frequency of environmental physical parameters and PUF response signals is increased to a preset multiple of the conventional mode.

[0104] Based on the environmental physical parameter data obtained by rapid sampling, the changing acceleration over multiple consecutive cycles is calculated and cached to form a real-time changing acceleration sequence.

[0105] Linear fitting is performed on the changing acceleration sequence. If the slope of the changing acceleration sequence of at least one environmental physical parameter is greater than zero, the situation level upgrade is triggered in advance, and a potential strong mutation early warning signal is generated to start the compensation vector loading preparation process.

[0106] This invention precisely classifies environmental conditions into three types: normal, non-stationary, and strong mutation, and sets differentiated processing strategies. Compensation vector loading is only activated when a strong mutation is detected, avoiding the need to initiate high-overhead correction processes for gradual or short-term fluctuations. This significantly reduces the system's computational load and resource consumption while ensuring authentication robustness.

[0107] The above description is merely an example and illustration of the structure of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the structure of the invention or exceed the scope defined by the present invention, they should all fall within the protection scope of the present invention.

Claims

1. A smart sensor based on physically unclonable functionality, characterized in that, include: The environmental parameter monitoring module monitors multiple environmental physical parameters of the PUF circuit and calculates the rate of change and acceleration of change of each environmental physical parameter within a predetermined time window. The environmental situation awareness module dynamically identifies the situation type of the current environment in which the PUF circuit is located based on the rate of change and the acceleration of change. The situation type is one of the following: normal situation, non-stationary situation, and strong sudden change situation. The compensation vector loading module, in response to a strong mutation situation signal, calls a specific compensation vector associated with the current strong mutation situation in the pre-stored compensation vector set, and applies the specific compensation vector to the response signal output terminal of the PUF circuit. The compensation effect verification module monitors the stability of the output response signal over multiple consecutive sampling periods after applying a specific compensation vector. If the stability meets the standard, it associates and stores the current strong change situation in the environment with the specific compensation vector; otherwise, it cancels the specific compensation vector and enables the standard correction process.

2. The intelligent sensor based on physically unclonable functionality according to claim 1, characterized in that, The dynamic identification process for environmental situation types includes: Set corresponding dynamic thresholds for the rate of change and acceleration of change of each environmental physical parameter; When the rate of change of all environmental physical parameters is less than their corresponding dynamic threshold of change, it is identified as a normal situation. Conversely, if the rate of change of any environmental physical parameter is greater than or equal to its corresponding dynamic threshold for rate of change: If the acceleration of all environmental physical parameters is less than their corresponding dynamic threshold for acceleration, it is identified as a non-stationary state. If the acceleration of change of at least one environmental physical parameter is greater than its corresponding dynamic threshold for acceleration change, it is identified as a strong abrupt change situation.

3. A smart sensor based on physically unclonable functionality according to claim 2, characterized in that, The process of setting the dynamic threshold includes: Continuously store the time series of each environmental physical parameter and its corresponding PUF response drift within the historical period; Extraction of the rate of change and acceleration sequences of various environmental physical parameters based on an adaptive sliding window; The standard deviation of the rate of change and acceleration of change of each environmental physical parameter, as well as the correlation coefficient between the change of each environmental physical parameter and the drift of the PUF response, are calculated in parallel. The normalized standard deviation is multiplied by the correlation coefficient, and then amplified by a configurable adjustment factor to generate a specific dynamic threshold for the rate of change and acceleration of change of each environmental physical parameter.

4. A smart sensor based on physically unclonable functionality according to claim 1, characterized in that, The process of constructing the pre-stored compensation vector set includes: In a controlled experimental environment, multiple preset environmental physical parameter mutation modes are applied sequentially. In each mutation mode, the raw response signal of the PUF circuit and its corresponding environmental physical parameter change rate and change acceleration data are collected; With the goal of achieving satisfactory response signals before and after compensation and minimizing the cumulative adjustment of compensation parameters, compensation parameters under each mutation mode are determined through iterative optimization. A mapping relationship is established between environmental mutation modes and their corresponding compensation parameters, and these relationships are stored to form a set of compensation vectors.

5. A smart sensor based on physically unclonable functionality according to claim 4, characterized in that, The specific compensation vector invocation process includes: Construct a first-order abrupt change situation feature matrix of the current environment, where each column of the matrix corresponds to an environmental physical parameter, and the column vector consists of the rate of change of the parameter and the over-threshold value of the acceleration of change. Transform each environmental mutation mode in the pre-stored compensation vector set into a sample matrix; Calculate the similarity between the first-order mutation situation matrix and each sample matrix, mark the sample matrix with the highest similarity as the candidate matrix, and use its corresponding compensation vector as the candidate compensation vector; The difference between the first-order mutation situation feature matrix and the candidate sample matrix is ​​calculated to obtain the environmental physical parameter deviation matrix; The deviation values ​​of each parameter in the environmental physical parameter deviation matrix are sorted to identify key environmental physical parameters whose deviation exceeds a set threshold. Based on the correlation characteristics between the key environmental physical parameters and the compensation vector, the candidate compensation vector is adjusted specifically to obtain the final specific compensation vector.

6. A smart sensor based on physically unclonable functionality according to claim 5, characterized in that, The parameter-specific adjustment of the candidate compensation vector includes: Based on the iterative optimization process, the proportional relationship between the historical change data of the key environmental physical parameters and the corresponding change of the compensation parameters is extracted. Based on the aforementioned proportional relationship and the actual deviation values ​​of the current key environmental physical parameters, calculate the adjustment component of the compensation vector; The adjustment component is superimposed with the corresponding parameter component in the candidate compensation vector to achieve directional fine-tuning of the candidate compensation vector.

7. A smart sensor based on physically unclonable functionality according to claim 1, characterized in that, The stability of the monitoring output response signal includes: Collect the output response signal of the PUF circuit for each sampling period and perform identity authentication; A time series analysis is performed on the identity authentication confidence scores over multiple consecutive sampling periods to determine whether the identity authentication confidence scores exhibit convergence characteristics within a preset authentication maintenance interval. If convergence characteristics are observed, the stability of the output response signal is confirmed to meet the standard.

8. A smart sensor based on physically unclonable functionality according to claim 7, characterized in that, The standard calibration procedure includes: Collect the raw response signal of the PUF circuit for each sampling period; Multiple raw response signals are merged and processed to obtain a composite response signal; Using the auxiliary data generated during the sensor registration phase and pre-stored locally, the integrated response signal is decoded and reconstructed to generate the reconstructed response signal to be verified. Perform identity authentication on the current reconstruction response signal to be verified, and obtain the current identity authentication confidence level; If the current identity authentication confidence level is at the high end of the preset authentication maintenance range, then identity authentication is completed.

9. A smart sensor based on physically unclonable functionality according to claim 1, characterized in that, It also includes the system self-optimization process: The verification results after each compensation vector loading and execution, including environmental strong mutation situation characteristic data, the called compensation vector identifier and stability verification results, are packaged together into an authentication event record; Based on the stability verification results, the authentication event records are classified and stored in different partitions of the historical case library, with records that meet the stability standard stored in the successful case partition and records that do not meet the stability standard stored in the failed case partition. Based on the authentication event records accumulated in the historical case library, perform at least one of the following optimization operations: The fluctuation characteristics of environmental parameters in the statistical failure case partitions are analyzed, and the adjustment coefficients in the dynamic threshold calculation are adjusted accordingly. Extract newly emerging environmental situation patterns from the successful case partitions and trigger incremental updates to the pre-stored compensation vector set.

10. A smart sensor based on physically unclonable functionality according to claim 1, characterized in that, In response to non-stationary situation signals, perform rapid sampling and situation prediction processing: Increase the sampling frequency of environmental physical parameters and PUF response signals to a preset multiple of the normal mode; Based on the environmental physical parameter data obtained by rapid sampling, the changing acceleration within multiple consecutive cycles is calculated and cached to form a real-time changing acceleration sequence. Linear fitting is performed on the changing acceleration sequence. If the slope of the changing acceleration sequence of at least one environmental physical parameter is greater than zero, the situation level upgrade is triggered in advance, and a potential strong mutation early warning signal is generated to start the compensation vector loading preparation process.

Images