以混沌系統為基礎之物聯網安全資料流

Secure Data Stream of IoT Based on Chaotic System

物聯網目前正蓬勃發展，物聯網資料流的安全性是資安的一個新課題。物聯網資料流的安全運算必須具有快速特性與有具良好的效率，而串流加密器具快速運算的優點，常應用於即時通訊的安全，其亦切合物聯網安全資料流的保密需求。混沌理論的輸出序列與初始值相關，其於密碼學之應用具有保密性、效率高、隨機性佳等優點，近來亦常見應用混沌理論於串流加密器的實現。本研究基於物聯網安全資料流的需求，結合不同低維度混沌理論建構金鑰流產生器的基本元件，以強化系統輸出序列的安全性。透過軟體實現產生輸出金鑰流，接著我們以FIPS PUB 140-1與美國國家科技標準局NIST的SP800對輸出金鑰流作亂度分析，結果顯示，在FIPS PUB 140-1的測試方面，過率為100％；在NISTSP 800的測試，金鑰流的通過率至少為92％。此外，我們進一步將此基於混沌系統的金鑰流產生器實現於物聯網系統，我們以物聯網平台Raspberry Pi為基礎，以實現一個以混沌系統為基礎的物聯網安全資料流，實作結果顯示在接收端可以解密得到正確的原始明文。

With the rapid development of IoT, the security of IoT is a new topic of information. The secure data stream of IoT requires the characteristics of operation fast, well efficiency, and that is just the advantage of stream cipher. The stream cipher is used for the security in real time communications, and it matches the requirement of the security of secure data stream of IoT. The outputs of chaotic system are highly related with the initial value, high randomness, and high efficiency. It is able to apply to cryptography. Recently the applications of chaotic theorem have been highlighted to enhance the security of stream cipher. In this paper, we combine 1-dementional chaotic systems with other basic elements of stream cipher and construct a hybrid chaotic stream cipher, to promote the period length, randomness, and the linear complexity of the output. After the implementation of stream cipher by Matlab, the output of the cipher will be examined via FIPS PUB 140-1 as well as NIST SP 800 for the randomness. For the test of FIPS PUB 140-1, all the pass rates of the proposed keystream generator are 100%. For the pass rates of NIST SP800-22, the proposed keystream generators is at least 92%. Besides, the proposed the hybrid chaotic based algorithm will be realized in the IoT platform Raspberry Pi for the security in IoT communication and wireless communication practically. From our implementation, the decrypted data is identical to the transmitted data correctly.