A Method and Device for Extracting Stable and Uniform Response Sequences for Storage-Class PUFs

Abstract

The present invention relates to a method and device for extracting a stable and uniform response sequence for a memory-based PUF, belonging to the technical field of PUF response sequence extraction, and solves the problem of poor stability of the response sequence in the prior art. The method for extracting a stable and uniform response sequence for a memory-based PUF provided by the present invention includes the following steps: performing multiple power-on measurements on the memory array of the memory-based PUF to obtain the power-on value and the marked value of each cell in the array; for each cell, if the power-on values of multiple power-on measurements are the same, the marked value is 1; otherwise, the marked value is 0; using the sliding window method, first grouping the cells with a marked value of 1, and then determining the stable cells based on the fractional Hamming weight of each group; sequentially extracting the power-on values of the stable cells according to the physical positions in the array to obtain the response sequence of the memory-based PUF. This method effectively improves the stability and uniformity of the PUF response sequence, thereby realizing the extraction of a stable and uniform response sequence for the memory-based PUF.

Description

Technical Field

[0001] The present invention relates to the technical field of PUF response sequence extraction, and in particular to a method and device for extracting a stable and uniform response sequence for a storage-type PUF. Background Art

[0002] Hardware security is an important branch in the field of information security and has received much attention and research from scholars at home and abroad. The Physical Unclonable Function (PUF) technology can generate output sequences with stability, uniqueness, and physical unclonability by utilizing the differences in physical devices or process characteristics, and is used in fields such as key generation, authorization, and authentication. Therefore, since its birth, it has attracted extensive attention from research institutions in the field of information security.

[0003] According to the differences in physical devices or process characteristics, PUF can be divided into non-electronic PUF and electronic PUF. With the update and replacement of physical devices and their production processes, electronic PUF has become the mainstream. And the PUF implemented based on an integrated circuit chip (also called silicon PUF, an electronic PUF) is currently the main object of theoretical research and practical application.

[0004] According to different implementation methods, silicon PUF can be divided into two categories: one is the PUF based on the inherent delay characteristics, including the arbiter PUF, the ring oscillator PUF, and the glitch PUF, etc.; the other is the PUF based on the memory, including the flip-flop PUF, the SRAM PUF, and the latch PUF, etc.

[0005] During the development and application process of PUF, three design indicators, namely "stability", "uniformity", and "uniqueness", are mainly concerned. Among them, "stability" refers to the similarity of the response sequences output by the same PUF under the same excitation at different times and environmental conditions, which is usually measured by the in-chip Hamming distance; "uniformity" refers to the proportion of 1s in the response sequences output by any PUF under various times and environmental conditions when excited, which is usually measured by the fractional Hamming weight (i.e., the number of 1s divided by the length of the response sequence), and ideally takes a value of 0.5; "uniqueness" refers to the difference degree of the response sequences output by different PUFs under the same excitation at the same time and environmental conditions, which is usually measured by the inter-chip Hamming distance.

[0006] As a well - studied implementation method of silicon PUF technology, storage - based PUF mainly relies on the manufacturing variables of the stable states of storage cell structures. The excitation is the address of a specific cell, and the response is the stable state of the cell. Due to uncontrollable factors in the manufacturing process, such as the uniformity of doping concentration and subtle differences in the length - to - width ratio of transistor channels, the stability of the response sequence output during actual operation is poor, making it difficult to meet the requirements of information security applications. Therefore, further processing is required to obtain a stable response sequence.

[0007] Typical methods for processing the response sequence of storage - based PUF, including the Temporal Majority Voting method, the Spatial Majority Voting method, and the Dark Bit Method, etc., have the following defects: They cannot ensure that the deviation of the fractional Hamming weight of the extracted response sequence from 0.5 is within an acceptable range, that is, they cannot ensure the uniformity of the extracted response sequence. Summary of the Invention

[0008] In view of the above analysis, the present invention aims to provide a method and device for extracting a stable and uniform response sequence for storage - based PUF to solve the problem of poor stability of the existing response sequence.

[0009] On the one hand, the present invention provides a method for extracting a stable and uniform response sequence for storage - based PUF, including the following steps:

[0010] Perform multiple power - on measurements on the memory array of the storage - based PUF to obtain the power - on value and the mark value of each cell in the array; wherein, for each cell, if the power - on values of the multiple power - on measurements are the same, the mark value is 1; otherwise, the mark value is 0;

[0011] Using the sliding window method, first group the cells with a mark value of 1, and then determine the stable cells based on the fractional Hamming weight of each group;

[0012] Extract the power - on values of the stable cells in sequence according to the physical positions in the array to obtain the response sequence of the storage - based PUF.

[0013] On the basis of the above - mentioned solution, the present invention has also made the following improvements:

[0014] Based on the further improvement of the above - mentioned method, let the total number of cells with a mark value of 1 be n1, and the number of bits of the response sequence to be extracted be m; before grouping, first determine the size relationship between n1 and m:

[0015] If n1 < m, reduce the number of power - on measurements of the array to re - determine n1, or reduce m until n1 ≥ m;

[0016] If n1≥m, the units with the marked value of 1 are grouped in the following way:

[0017] Extract the units with the marked value of 1 in the array in sequence according to the physical positions in the array; and group the extracted units with the marked value of 1 in a way that every m units form a group, where: the starting unit of the i-th group is the i-th unit with the marked value of 1, and the ending unit is the (i + m - 1)-th unit with the marked value of 1, and i takes values from 1, ……, n1 - m + 1.

[0018] Based on a further improvement of the above method, if n1 = m, there is only one group, and directly take the m units in this group as stable units.

[0019] Based on a further improvement of the above method, if n1>m, there are multiple groups, and stable units are obtained by performing the following operations:

[0020] Calculate the fractional Hamming weight FHW1 of the first group where i = 1, and let the optimal fractional Hamming weight

[0021] FHW_opt = FHW1;

[0022] Judge whether n1 - i≥m holds.

[0023] If it does not hold, take the first m units with the marked value of 1 as stable units;

[0024] If it holds, then i = i + 1, and calculate the fractional Hamming weight FHW of the i-th group i , and judge whether FHW i satisfies |FHW i - 0.5| < |FHW_opt - 0.5|.

[0025] If it satisfies, let FHW_opt = FHW i , re-mark the first i - 1 units with the marked value of 1 as 0, and jump to judge whether n1 - i≥m holds;

[0026] If it does not satisfy, directly jump to judge whether n1 - i≥m holds.

[0027] Based on a further improvement of the above method, set the range of the fractional Hamming weight threshold;

[0028] If n1 = m, there is only one group, calculate the fractional Hamming weight of this group, if the fractional Hamming weight of this group meets the requirements of the fractional Hamming weight threshold range, then take the m units in this group as stable units; otherwise, reduce the number of power-on measurements of the array to re-determine n1, or reduce m until n1≥m.

[0029] Based on further improvements to the above method, if n1 > m, stable units are obtained by performing the following operations:

[0030] Calculate the fractional Hamming weight of each group in sequence, and use m units in the first group that meets the requirements of the fractional Hamming weight threshold range as stable units.

[0031] Based on further improvements to the above method, the fractional Hamming weight threshold range is [0.5 - FHWTh, 0.5 + FHWTh]; where FHWTh represents the fractional Hamming weight threshold parameter.

[0032] Based on further improvements to the above method, the fractional Hamming weight threshold parameter FHWTh ∈ [0, η], 0 < η << 0.5.

[0033] Based on further improvements to the above method, the fractional Hamming weight refers to the ratio of the sum of the power-on values of all units in the group to the unit length of the group.

[0034] On the other hand, the present invention also provides a device for extracting a stable and uniform response sequence for a storage-class PUF, including:

[0035] A stable bit marker, which is used to perform multiple power-on measurements on the memory array of the storage-class PUF to obtain the power-on value and marker value of each unit in the array; where for each unit, if the power-on values of the multiple power-on measurements are the same, the marker value is 1; otherwise, the marker value is 0;

[0036] A stable bit selector, which is used to use the sliding window method to first group the units with a marker value of 1, and then determine the stable units based on the fractional Hamming weight of each group;

[0037] A stable bit outputter, which is used to sequentially extract the power-on values of the stable units according to the physical positions in the array to obtain the response sequence of the storage-class PUF.

[0038] Compared with the prior art, the present invention can achieve at least one of the following beneficial effects:

[0039] The method for extracting a stable and uniform response sequence for a storage-class PUF provided by the present invention first uses the stable bit marking method to obtain n1 cells with a marked value of 1, which is used to ensure the stability of the response sequences corresponding to the stable cells in the stable group; then, following the principle of optimal fractional Hamming weight or fractional Hamming weight meeting the requirements of a set threshold range within linear time, m cells are selected from the n1 cells with a marked value of 1 as stable cells, which is used to ensure the uniformity of the response sequences corresponding to the stable cells. That is, on the premise of ensuring the stability of the PUF response sequence, the uniformity of the PUF response sequence is improved, thereby realizing the extraction of a stable and uniform response sequence for the storage-class PUF. This method can ensure that for this type of PUF, under various environmental conditions such as temperature and aging, stable and uniform PUF response sequences can be accurately generated using the vast majority of integrated circuit processes.

[0040] In the present invention, the above technical solutions can also be combined with each other to achieve more preferred combination solutions. Other features and advantages of the present invention will be described in the subsequent description, and some advantages will become obvious from the description or can be understood by implementing the present invention. The objectives and other advantages of the present invention can be achieved and obtained through the content specifically pointed out in the description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The drawings are only for the purpose of showing specific embodiments and are not considered to be a limitation of the present invention. Throughout the drawings, the same reference signs represent the same components.

[0042] Figure 1 is a flowchart of the method for extracting a stable and uniform response sequence for a storage-class PUF provided by the present invention;

[0043] Figure 2 is a schematic diagram of the sliding window method provided by the present invention;

[0044] Figure 3 is a schematic structural diagram of the device for extracting a stable and uniform response sequence for a storage-class PUF provided by the present invention;

[0045] Reference Signs:

[0046] 1 - Stable bit marker in the registration stage, 2 - Stable bit selector in the registration stage, 3 - Stable bit outputter. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0047] The following will specifically describe the preferred embodiments of the present invention in conjunction with the drawings. The drawings form a part of this application and are used together with the embodiments of the present invention to explain the principle of the present invention, rather than to limit the scope of the present invention.

[0048] Embodiment 1

[0049] Specific Embodiment 1 of the present invention discloses a method for extracting a stable and uniform response sequence for a storage-type PUF. A storage-type PUF generally includes a memory array. The flowchart of this method is as shown in Figure 1 and includes the following steps:

[0050] Step S1: Perform multiple power-on measurements on the memory array of the storage-type PUF to obtain the power-on value and the mark value of each cell in the array;

[0051] For each cell, during multiple power-on measurements, if there is a flip in the power-on value of one measurement compared to the power-on value of the first measurement, the mark value is 0, which is called an unstable bit; otherwise, the mark value is 1, which is called a stable bit; the mark value is stored as auxiliary data. That is, for each cell, if the power-on values of the multiple power-on measurements are the same, the mark value is 1; otherwise, the mark value is 0.

[0052] Specifically, the number of power-on measurements can be specifically determined based on the test accuracy and experience, and this embodiment does not limit it; the power-on value refers to the value measured after the storage-type PUF is powered on; according to the design principle of the storage-type PUF, as long as it is powered on, the power-on value 0 or 1 can be measured; for a certain cell, if the power-on value measured when it is powered on for the first time is 1 and the power-on value measured when it is powered on the next time is 0, it is considered a flip; vice versa.

[0053] Step S2: Use the sliding window method to first group the cells with a mark value of 1, and then determine the stable cells based on the fractional Hamming weight of each group; where the fractional Hamming weight refers to the ratio of the sum of the power-on values of all cells in the group to the length of the group of cells.

[0054] Specifically, in this embodiment, the process of determining the stable cells based on the sliding window method includes:

[0055] Step S21: Group the cells with a mark value of 1:

[0056] Assume that the total number of cells with a mark value of 1 is n1, and the number of bits of the response sequence to be extracted is m; before grouping the cells with a mark value of 1, first determine the size relationship between n1 and m:

[0057] (1) If n1 < m, reduce the number of power-on measurements of the array to re-determine n1, or reduce m until n1 ≥ m; it should be noted that for a well-designed storage-type PUF, by appropriately adjusting the number of power-on measurements or the number of bits of the response sequence to be extracted, n1 ≥ m can be achieved.

[0058] (2) If n1 ≥ m, group the cells with a mark value of 1 in the following way:

[0059] Extract the cells with a marked value of 1 in the array in sequence according to their physical positions in the array; and group the extracted cells with a marked value of 1 in a way that every m cells form a group, where: the starting cell of the i-th group is the i-th cell with a marked value of 1, and the ending cell is the (i + m - 1)-th cell with a marked value of 1, and i takes values from 1 to n1 - m + 1. At this time, n1 ∈ [m, n].

[0060] Step S22: Determine stable cells:

[0061] After determining the grouping, m cells can be selected from the n1 cells with a marked value of 1 as stable cells according to the principle of optimal fractional Hamming weight within linear time or the requirement that the fractional Hamming weight meets the set threshold range, so as to ensure the uniformity of the response sequence corresponding to the stable cells.

[0062] Implementation method 1:

[0063] When following the principle of optimal fractional Hamming weight within linear time, the stable cells are selected by performing the following operations:

[0064] (1) If n1 = m, there is only one group. Preferably, at this time, the m cells in this group can be directly used as stable cells;

[0065] (2) If n1 > m, there are multiple groups. Optionally, the following operations are performed to obtain the group with the optimal fractional Hamming weight, and the m cells in the optimal group are used as stable cells; specifically:

[0066] Calculate the fractional Hamming weight FHW1 of the i = 1 group, and let the optimal fractional Hamming weight

[0067] FHW_opt = FHW1;

[0068] Judge whether n1 - i ≥ m holds,

[0069] If it does not hold, use the first m cells with a marked value of 1 as stable cells;

[0070] If it holds, then i = i + 1, and calculate the fractional Hamming weight FHW i , and judge whether FHW i satisfies |FHW i - 0.5| < |FHW_opt - 0.5|,

[0071] If it satisfies, let FHW_opt = FHW i , re - mark the first i - 1 cells with a marked value of 1 as 0, and jump to judge whether n1 - i ≥ m holds;

[0072] If not satisfied, directly jump to determine whether n1 - i ≥ m holds.

[0073] After processing by the sliding window method (as Figure 2 shown), at this time, the number of units with a marker value of 1 becomes n2, and n2 ∈ [m, n1].

[0074] The principle of selecting stable units is as follows: Since 0.5 indicates that the number of units with a marker value of 0 is the same as the number of units with a marker value of 1; the smaller the difference between the fractional Hamming weight and 0.5, the better the uniformity of the response sequence. By calculating |FHW i - 0.5| < |FHW_opt - 0.5|, continuously search for and optimize the optimal fractional Hamming weight FHW_opt in linear time, and determine the final stable units based on the optimal fractional Hamming weight, thus ensuring the uniformity of the response sequence.

[0075] Implementation method 2:

[0076] When following the principle that the fractional Hamming weight meets the requirements of the set threshold range, first set the fractional Hamming weight threshold range, and select stable units by performing the following operations:

[0077] (1) If n1 = m, there is only one group. Calculate the fractional Hamming weight of this group. If the fractional Hamming weight of this group meets the requirements of the fractional Hamming weight threshold range, then take the m units in this group as stable units; otherwise, reduce the number of power - on measurements of the array to re - determine n1, or reduce m. For a well - designed memory - based PUF and a reasonably set fractional Hamming weight threshold range, by appropriately adjusting the number of power - on measurements or the number of response sequence bits m to be extracted, stable units that meet the requirements can be found.

[0078] (2) If n1 > m, obtain the stable units by performing the following operations:

[0079] Calculate the fractional Hamming weight of each group in turn, and take the m units in the first group that meets the requirements of the fractional Hamming weight threshold range as stable units.

[0080] Among them, the fractional Hamming weight threshold range is [0.5 - FHWTh, 0.5 + FHWTh]; where FHWTh represents the fractional Hamming weight threshold parameter. The fractional Hamming weight threshold parameter FHWTh ∈ [0, η], 0 < η << 0.5.

[0081] For the case where n1 > m, Implementation Method 1 selects the final stable cell by finding the optimal fractional Hamming weight within linear time; while Implementation Method 2 sets a reasonable threshold and takes the first grouped cell that meets the requirements of the fractional Hamming weight threshold range as the stable cell; therefore, compared with Implementation Method 1, Implementation Method 2 can shorten the search time.

[0082] It should be noted that the number of response sequence bits to be extracted is specifically determined according to the use of the output response sequence; exemplarily, when the output response sequence is used for key generation, the number of response sequence bits can be determined according to the length requirement of the key.

[0083] Step S3: Extract the power-on values of the stable cells in sequence according to the physical positions in the array to obtain the response sequence of the memory-class PUF.

[0084] Compared with the prior art, the method for extracting a stable and uniform response sequence for a memory-class PUF provided in this embodiment first uses the stable bit marking method to obtain n1 cells marked as 1 to ensure the stability of the response sequence corresponding to the stable cells in the stable group; then follows the principle of optimal fractional Hamming weight or fractional Hamming weight meeting the set threshold within linear time to select m cells from the n1 cells marked as 1 as the stable group to ensure the uniformity of the response sequence corresponding to the stable cells in the stable group. That is, on the premise of ensuring the stability of the PUF response sequence, the uniformity of the PUF response sequence is improved, thereby realizing the extraction of a stable and uniform response sequence for the memory-class PUF. This method can ensure that this type of PUF can accurately generate a stable and uniform PUF response sequence under various environmental conditions such as temperature and aging using the vast majority of integrated circuit processes.

[0085] Embodiment 2

[0086] Specific Embodiment 2 of the present invention discloses a device for extracting a stable and uniform response sequence for a memory-class PUF, and the structural schematic diagram is as Figure 3 shown. The device includes:

[0087] A stable bit marker 1 for performing multiple power-on measurements on the memory array of the memory-class PUF to obtain the power-on value and marker value of each cell in the array; wherein, for each cell, if the power-on values of the multiple power-on measurements are the same, the marker value is 1; otherwise, the marker value is 0;

[0088] A stable bit selector 2 for using the sliding window method to first group the cells with a marker value of 1, and then select stable cells based on the fractional Hamming weight of each group;

[0089] The stable bit outputter 3 is configured to sequentially extract the power-on values of the stable cells according to the physical positions in the array, so as to obtain the response sequence of the memory type PUF.

[0090] For the specific implementation process of the device embodiment of the present invention, reference may be made to the above method embodiment, and details are not described herein again.

[0091] Since the device embodiment of the present invention has the same principle as the above method embodiment, the device also has the corresponding technical effects of the above method embodiment.

[0092] Embodiment 3

[0093] In the specific Embodiment 3 of the present invention, taking the extraction of a stable and uniform PUF response sequence from an 1120-bit SRAM memory array as an example, the specific implementation process of the above embodiment is described as follows:

[0094] The extraction process refers to Figures 1-3 : It includes the stable bit marker 1 in the registration stage, the stable bit selector 2 in the registration stage, and the stable bit outputter 3. According to the given number of power-on measurements t = 13 and the number of bits m = 160 of the response sequence to be extracted, first, in the registration stage, 13 power-on values generated during the power-on and power-off processes of the 1120-bit SRAM memory array are collected and measured. The stable bit marker 1 marks more than 160 stable cells (i.e., cells with a marked value of 1), and then the sliding window method of the stable bit selector 2 (refer to Figure 2 ) is used to select 160 consecutive cells among the stable bits measured 13 times as the extracted stable group. Each cell in the stable group is a stable cell (also called a "stable response cell"); finally, the stable bit outputter 3 uses the power-on values of the first 160 stable cells with a marked value of 1 as the stable and uniform response sequence.

[0095] The stable bit marker 1 in the registration stage is mainly used to monitor the bit cells in the 1120-bit SRAM memory array, perform 13 power-on and power-off measurements, and collect the response output results 1 or 0 each time. If all are 1 or 0, the marked value of the bit cell is set to 1, otherwise the marked value is set to 0. The number of cells with a marked value of 1 is 786.

[0096] The stable bit selector 2 in the registration stage is mainly used to select 786 bit cells with a marked value of 1 in the 1120-bit SRAM memory array. With the goal of being optimal within the linear time of the registered value fractional Hamming weight, the sliding window method is used to select 160 consecutive bit cells with a marked value of 1 as the extracted stable response cells, and the bit cells with a marked value of 1 before the first selected cell are re-marked as 0. After being processed by the sliding window method, the number of marked values of 1 is reduced to 694. Among them, the specific steps of the sliding window method can be referred to the relevant content in Embodiment 1. Thus, 694 bit cells with a marked value of 1 in the 1120-bit SRAM memory array are obtained, and the stable bit outputter 3 is used to output the power-on values of the first 160 bit cells with a marked value of 1 as the extracted stable uniform response sequence.

[0097] Those skilled in the art can understand that all or part of the processes of implementing the methods in the above embodiments can be completed by instructing relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. Among them, the computer-readable storage medium is a disk, an optical disc, a read-only memory, or a random access memory, etc.

[0098] The above is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present invention should be covered by the protection scope of the present invention.

Claims

1. A method for extracting a stable and uniform response sequence for a storage-class PUF, characterized in that, It includes the following steps: Perform multiple power-on measurements on the memory array of the storage-class PUF to obtain the power-on values and tag values of each cell in the array; wherein, for each cell, if the power-on values of the multiple power-on measurements are the same, the tag value is 1; otherwise, the tag value is 0; Using the sliding window method, first group the cells with a tag value of 1, and then determine the stable cells based on the fractional Hamming weight of each group; Extract the power-on values of the stable cells in sequence according to the physical positions in the array to obtain the response sequence of the storage-class PUF; The fractional Hamming weight refers to the ratio of the sum of the power-on values of all cells in the group to the cell length of the group.

2. The method for extracting a stable and uniform response sequence for a storage-class PUF according to claim 1, wherein Let the total number of cells with a tag value of 1 be n1, and the number of bits of the response sequence to be extracted be m; before grouping the cells with a tag value of 1, first determine the size relationship between n1 and m: If n1 < m, reduce the number of power-on measurements on the array to re-determine n1, or reduce m until n1 ≥ m; If n1 ≥ m, group the cells with a tag value of 1 in the following way: Extract the cells with a tag value of 1 in the array in sequence according to the physical positions in the array; and group the extracted cells with a tag value of 1 in groups of every m cells, where: the starting cell of the i-th group is the i-th cell with a tag value of 1, and the ending cell is the (i + m - 1)-th cell with a tag value of 1, and i takes 1, ……, n1 - m + 1.

3. The method for extracting a stable and uniform response sequence for a storage-class PUF according to claim 2, wherein If n1 = m, there is only one group, and directly use the m cells in this group as stable cells.

4. The method for extracting a stable and uniform response sequence for a storage-class PUF according to claim 3, wherein If n1 > m, there are multiple groups, and obtain stable cells by performing the following operations: Calculate the fractional Hamming weight FHW1 of the i = 1 group, and let the optimal fractional Hamming weight FHW_opt = FHW1; Judge whether n1 - i ≥ m holds, If it does not hold, use the first m cells with a tag value of 1 as stable cells; If it holds, then i = i + 1, and calculate the fractional Hamming weight FHW of the i-th group i , and judge FHW i whether it satisfies |FHW i - 0.5| < |FHW_opt - 0.5| If satisfied, let FHW_opt = FHW i , relabel the first i - 1 cells with a marked value of 1 as 0, and jump to determine whether n1 - i ≥ m holds; If not satisfied, directly jump to judge whether n1 - i ≥ m holds.

5. The method for extracting a stable and uniform response sequence for a storage-class PUF according to claim 2, wherein Set the range of the fractional Hamming weight threshold; If n1 = m, there is only one group, calculate the fractional Hamming weight of this group, if the fractional Hamming weight of this group meets the requirements of the fractional Hamming weight threshold range, use the m cells in this group as stable cells; otherwise, reduce the number of power-on measurements on the array to re-determine n1, or reduce m until n1 ≥ m.

6. The method for extracting a stable and uniform response sequence for a storage-class PUF according to claim 5, wherein If n1 > m, obtain stable cells by performing the following operations: Calculate the fractional Hamming weight of each group in sequence, and use the m cells in the first group that meet the requirements of the fractional Hamming weight threshold range as stable cells.

7. The method for extracting a stable and uniform response sequence for a storage-class PUF according to claim 5 or 6, characterized in that, The fractional Hamming weight threshold range is [0.5 - FHWTh, 0.5 + FHWTh]; where FHWTh represents the fractional Hamming weight threshold parameter.

8. The method for extracting a stable and uniform response sequence for a storage-class PUF according to claim 7, characterized in that, The fractional Hamming weight threshold parameter FHWTh ∈ [0, η], 0 < η << 0.

5.

9. A stable and uniform response sequence extraction device for a storage-class PUF, characterized in that It includes: A stable bit marker, which is used to perform multiple power-on measurements on the memory array of the storage-class PUF to obtain the power-on value and marker value of each cell in the array; for each cell, if the power-on values of the multiple power-on measurements are the same, the marker value is 1; Otherwise, the marker value is 0; A stable bit selector, which is used to group the cells with a marker value of 1 by using the sliding window method, and then determine the stable cells based on the fractional Hamming weight of each group; A stable bit outputter, which is used to sequentially extract the power-on values of the stable cells according to the physical positions in the array to obtain the response sequence of the storage-class PUF; The fractional Hamming weight refers to the ratio of the sum of the power-on values of all cells in the group to the cell length of the group.

Images