National Institute of Advanced Industrial Science and Technology

Toshihiro Matsui

Kazukuni Kobara

Toshihiro Katashita

Yohei Hori

A method and a program capable of controlling an error rate of device-specific information are provided. Provided is the method for controlling an error rate of device-specific information, including a step Sof: inputting each of i (i is an arbitrary natural number) challenges, j times (j is an arbitrary natural number), into a PUF mounted chip; leaving j responses intact (j'=j) or processing j responses into j&#x2032; pieces (0&#x3c;j&#x2032;&#x3c;j); and registering them in the database beforehand in association with each piece of the input data, a step Sof inputting i challenges into the database, a step Sof: reading j&#x2032; responses corresponding to the respective i challenges from the database; concatenating the j&#x2032; responses for each piece of the input data; further concatenating the concatenated data by k&#x2032; repetitions (0&#x3c;k&#x2032;&#x2264;k, and k is an arbitrary natural number, but is a natural number of 2 or more if the i and the j are both 1); obtaining the concatenated (j&#x2032;&#xd7;k&#x2032;) responses for each piece of the input data; and further concatenating them also for different input data to obtain concatenated (i&#xd7;j&#x2032;&#xd7;k&#x2032;) responses and thereby generate reference data, a step Sof: inputting i challenges, k times, for each challenge into the PUF mounted chip; leaving obtained k responses intact as k&#x2032;=k or processing the obtained k responses into k&#x2032; pieces (0&#x3c;k&#x2032;&#x3c;k); concatenating obtained k&#x2032; responses by j&#x2032; repetitions for each response; further concatenating them for all of the k&#x2032; responses; further concatenating concatenated (j&#x2032;&#xd7;k&#x2032;) responses also for different input data; and obtaining concatenated (i&#xd7;j&#x2032;&#xd7;k&#x2032;) responses to generate synthesized output data, and a step Sof deciding whether or not the synthesized output data matches the reference data (specifically, whether a Hamming distance between both data is a threshold value or less), and the method determines whether or not the error rate of the synthesized output data is within a preset range based on the decision result in step S, and changes at least one of i, j, j&#x2032;, k, and k&#x2032; to repeat steps Sto Suntil the error rate falls within the preset range if the error rate is determined not to be within the preset range.

Growing community of inventors

Tsukuba, Japan

Toshihiro Matsui