Smart hardware security engine using biometric features and hardware-specific features

Patent metadata

Field	Value
Publication number	US-11698952-B2
Application number	US-202016864902-A
Country	US
Kind code	B2
Filing date	May 1, 2020
Priority date	May 2, 2019
Publication date	Jul 11, 2023
Grant date	Jul 11, 2023

Abstract

A smart hardware security engine using biometric features and hardware-specific features is provided. The smart security engine can combine one or more entropy sources, including individually distinguishable biometric features, and hardware-specific features to perform secret key generation for user registration and authentication. Such hybrid signatures may be distinct from person-to-person (e.g., due to the biometric features) and from device-to-device (e.g., due to the hardware-specific features) while varying over time. Thus, embodiments described herein can be used for personal device authentication as well as secret random key generation, significantly reducing the scope of an attack.

First claim

What is claimed is:

1. A method for biometric authentication, the method comprising:
generating a first biometric feature vector from biometric data, the first biometric feature vector identifying a user;
generating biometric variability data based on the biometric data;
generating a hardware-specific feature vector;
concatenating the first biometric feature vector with the biometric variability data to generate a second biometric feature vector; and
generating a secret key based on the second biometric feature vector and the hardware-specific feature vector.

2. The method of claim 1 , wherein generating the first biometric feature vector comprises producing a multi-bit randomized number from the biometric data which differentiates between different users.

3. The method of claim 2 , wherein generating the first biometric feature vector further comprises applying a learning algorithm trained to generate the multi-bit randomized number unique to each of the different users.

4. The method of claim 1 , further comprising receiving the biometric data through a biometric sensor.

5. The method of claim 4 , wherein:
the biometric data comprises cardiac data;
the biometric variability data is an averaged heart rate variability value; and
the biometric sensor comprises at least one of an electrocardiography (ECG) sensor or a photoplethysmography (PPG) sensor.

6. The method of claim 1 , wherein the hardware-specific feature vector comprises a physical unclonable function (PUF) vector.

7. The method of claim 1 , wherein the secret key is generated to register the user as a registered user.

8. The method of claim 1 , wherein the secret key is generated to authenticate a registered user.

9. The method of claim 8 , wherein when authenticating the registered user, the first biometric feature vector is extracted over fewer samples of the biometric data than during registration.

10. A circuit, comprising:
hardware-specific feature circuitry capable of producing a hardware-specific feature vector; and
a processor configured to:
receive a biometric signal;
generate biometric variability data based on the biometric signal;
extract a first biometric feature vector from the biometric signal, the first biometric feature vector identifying a user;
concatenating the first biometric feature vector with the biometric variability data to generate a second biometric feature vector; and
generate a secret key based on the second biometric feature vector and the hardware-specific feature vector from the hardware-specific feature circuitry.

11. The circuit of claim 10 , wherein the processor comprises a learning algorithm implemented on a neural network to extract the first biometric feature vector.

12. The circuit of claim 11 , wherein the neural network is capable of differentiating between biometric signals of different users.

13. The circuit of claim 12 , wherein the neural network comprises at least one hidden layer for extracting the first biometric feature vector.

14. The circuit of claim 13 , wherein the at least one hidden layer is a fully connected layer.

15. The circuit of claim 13 , wherein the at least one hidden layer is a sparsely connected layer.

16. The circuit of claim 13 , wherein:
the neural network does not have an output layer; and
the first biometric feature vector is based on values of a last layer of the at least one hidden layer.

17. The circuit of claim 16 , wherein the neural network is trained by minimizing an application-specific cost function formulated as:

cost
=

-

μ
intra

-

μ
inter

σ
intra

+

σ
inter

wherein μ intra is a mean of an intra-subject cosine similarity distribution, μ inter is a mean of an inter-subject cosine similarity distribution, σ intra is a standard deviation of the intra-subject cosine similarity distribution, and σ inter is a standard deviation of the inter-subject cosine similarity distribution.

18. The circuit of claim 10 , wherein the hardware-specific feature circuitry comprises static random-access memory (SRAM) configured to generate a physical unclonable function (PUF) vector.

19. A device, comprising:
a memory; and
a processor coupled to the memory and configured to:
receive biometric data;
generate biometric variability data;
extract a first biometric feature vector unique to a user from the biometric data;
generate a hardware-specific feature vector;
concatenate the first biometric feature vector with the biometric variability data to generate a second biometric feature vector; and
authenticate the user based on the second biometric feature vector and the hardware-specific feature vector.

20. The device of claim 19 , wherein the device is a wearable device.

21. The device of claim 20 , wherein:
the wearable device comprises a cardiac sensor configured to provide the biometric data of the user; and
the biometric variability data represents heart rate variability.

22. The device of claim 19 , wherein the hardware-specific feature vector comprises a physical unclonable function (PUF) vector.

Assignees

• Seo Jae Sun
• Yin Shihui
• Cherupally Sai Kiran
• Univ Arizona State

Inventors

• Seo Jae-Sun
• Yin Shihui
• Cherupally Sai Kiran

Classifications

• H04L9/0866

  involving user or device identifiers, e.g. serial number, physical or biometrical information, DNA, hand-signature or measurable physical characteristics · CPC title

• H04L9/0869

  involving random numbers or seeds · CPC title

• G06F21/32 Primary

  using biometric data, e.g. fingerprints, iris scans or voiceprints · CPC title

• H04L9/3278

  using physically unclonable functions [PUF] · CPC title

• G06N3/08

  Learning methods · CPC title