Skip to main content

Cooperative Communications with Confidential Messages

Final Report Summary - CCCM (Cooperative Communications with Confidential Messages)

The main research objective of this project is to explore the possibility of cooperation of multiple terminals in a wireless communication network in order to transmit and receive data that is protected from unwanted users. All wireless systems have a broadcast character and while transmitting data to a legitimate recipient there is a possibility that other “unwanted participants” intercept the transmitted information. In classical systems the security is solved by the exchange of cryptographic keys over dedicated channels that are assumed to be perfectly secure. Future wireless systems will consists of various heterogeneous nodes for which it will be difficult or impossible to exchange keys while guaranteeing perfect secrecy. Therefore, the idea is to design a system in which the data is transmitted with minimal power such that malicious nodes will have a little or no possibility to receive the data that is not intended to them. In order to increase the throughput and security level between the source and the destination node, other friendly users can cooperate with the sender. For a given power, we aim at maximizing the reliability and the throughput of wireless ad-hoc network while the communication stays confidential.

During the outgoing phase of the project, we studied various scenarios for cooperative communications with confidential messages. A typical system consists of two legitimate nodes (one sender and one receiver), one eavesdropper and one or more cooperating (helping) nodes. We considered various channel models, and in particular we studied the role of the cooperating nodes. Main conclusions after this phase are that the role of the cooperating nodes is extremely important in improving the overall security of a wireless network. One way of helping is by receiving the data from the sender and re-sending (relaying) them to the receiver, hence increasing the quality of communication between the two legitimate nodes. Another way of helping is to send interference to the eavesdropper, and therefore, reducing the quality of the channel between the sender and the eavesdropper. The former is called cooperation by relaying and the latter cooperation by jamming. Note also that cooperating nodes can be either mobile or fixed, which allows the use of the existing infrastructure to enhance the secure transmission. That means hybrid systems may be deployed and they even show better results. The main results show that cooperating nodes increase not only the secrecy rate but also the so-called secrecy rate region. It was also concluded that it is possible to have cooperating nodes that do not need to have access to the transmitted confidential information. They are called blind cooperating nodes. The blind helper concept is quite significant since it can be useful in many realistic scenarios, for example transmitting confidential information in a network where the cooperating nodes are untrusted.

During the return phase of the project, we studied the effect of the cooperating (helping) nodes, the possibility of combining the information-theoretic secrecy and cryptographic secrecy, their implementation and comparison. We proposed a scheme in which perfectly secure links that are supported by the helpers are used for exchanging secret keys between the legitimate parties. Once the secret key is exchanged, it is used to encrypt the confidential data. If the secret key changes in some regular periods, it is possible to obtain a system with high level of security. This is a great example where the best notion of secrecy, namely, information-theoretic secrecy complements the conventional cryptographic secrecy. It allows getting much better throughput for the secure data. The throughput is studied in terms of the probability that the secrecy rate is non-vanishing. In this system that combines information-theoretic and cryptographic secrecy, we study four scenarios in which both cooperating nodes and eavesdropping nodes are positioned according to two dimensional Poisson distribution and uniform distribution respectively. Finally, we concluded this interesting project with considering several practical scenarios where cooperative information-theoretic secrecy systems can be used.

The main results are the following:
- It is better to use cooperating nodes as jammers than relays,
- In average, it is more efficient to select the “strongest” cooperating node than using all available cooperating nodes. Even in this case, choosing the best jammer is better than choosing the best relay,
- If the cooperating nodes are untrusted, then it is straightforward to use them as jammers. They can also be used as relays, however the benefit will be smaller than in the case they are trusted,
- In the case where information-theoretic secrecy is used for the exchange of the cryptographic key (which is then used to encrypt the data), the results are much better,
- When the cooperating and eavesdropping nodes are randomly distributed we have four different scenarios (each group is either distributed according to the uniform or Poisson distribution) that behave quite similarly
- In the case where the cooperating nodes are positioned in fixed locations, the throughput in terms of non-vanishing probability of the secrecy rate is even better. We consider two different positioning systems for the cooperating nodes: Hexagonal lattice and square lattice. When the cooperating nodes are positioned within the square lattice the results are slightly better. That comes from the fact that the helping nodes are more regularly distributed than in the mobile (random) case. In both scenarios eavesdroppers are mobile, i.e. randomly positioned.

We also started to look at three novel approaches in information theory that may be linked to information-theoretic secrecy and cooperation:
1. The non-asymptotic approach for cooperative communication systems with confidential messages (with Prof. Sergio Verdú, Shannon Award Winner 2007),
2. Use of polar codes for cooperative communication systems with confidential messages (with Prof. Shlomo Shamai, Shannon Award Winner 2011),
3. The I-MMSE (mutual information versus minimum mean square error) approach for cooperative communication systems with confidential messages (with both Profs. Sergio Verdú and Shlomo Shamai).
The latest three approaches are quite fundamental and can lead to new understanding of the future reliable and secure communication systems. The fellow will continue to work and produce results in these areas.

As for the practical use of the information-theoretic security, there are several promising scenarios that can have impact in practice: mobile ad-hoc networks, near field and RFID networks, network coding systems, magnetic recording, privacy in the online social networks, privacy in the electric utility network (smart grids), genomic privacy (privacy of the DNA data), security in vehicle-to-vehicle communication and supply chain security.