Modes of operation for Compressed Sensing based encryption

Compressed Sensing offers the possibility to jointly compress and encrypt sparse or compressible signals during the sampling process, i.e. directly at the sensor level. For the purpose of encryption, only the sampling matrix, which is also needed for the signal recovery, must be kept secret. However, this type of encryption scheme is no longer secure if multiple signals are encrypted, since the same plaintext will always yield the same ciphertext. This thesis proposes modes of operations for Compressed Sensing that allow the secure encryption of multiple signals.
First, different attacks and threat models are analyzed in order to develop security definitions for Compressed Sensing based encryption and Compressed Sensing modes of operations. The results from this analysis are used to design a general model for Compressed Sensing modes, which ensure confidentiality and additionally reduce the information leakage of Compressed Sensing encryption. The security and performance of cryptographic constructions that are suitable for implementing the general model are evaluated and three dedicated modes with different quality of service properties are derived from the general design. These modes of operation and their corresponding security analysis are the first main result of this thesis.
In addition to confidentiality, which is achieved through encryption, another important security service is data integrity. This work also examines so-called authenticated-encryption modes for Compressed Sensing, which ensure confidentiality and data integrity simultaneously. A general model for Compressed Sensing with authenticated-encryption is developed and dedicated schemes are derived from this model. The security of these schemes is reduced to the security of the underlying cryptographic constructions and primitives. Compressed Sensing with authenticated encryption provides security even in the case where an adversary has the ability to tamper with the ciphertext.
The final contribution of this thesis is the integration of the proposed Compressed Sensing modes into a distributed application. A software architecture for Industry 4.0 applications is developed and implemented. The designed system includes all necessary components for the usage of Compressed Sensing modes, like key-establishment and parameter exchange. Next to security, a main design goal of this system is usability. The user just determines the security service for his application and all security relevant operations are handled by the software, transparent for the user. This approach prevents security problems that are caused by inappropriate usage of cryptographic schemes.
The Compressed Sensing modes developed in this thesis are suitable for a wide range of real world applications due their joint sampling, compression and end-to-end security that starts at the sensor level. The proposed software-system completes the contribution of this work, since it realizes all necessary components for the usage of the designed modes.