| dc.description.abstracteng | The main objective of this dissertation is to investigate security solutions and issues at the lower layers in wireless networks.
In the first part, the potential of the physical layer in providing security solutions is investigated. Recently, it has been found that the multipath wireless channel in TDD wireless communications can provide a common reciprocal source of randomness that can be leveraged in secret key generation and agreement. Based on this property, many key generation mechanisms have been proposed. In contrary to the common direct quantization and extraction mechanisms, we propose two intelligent mechanisms for secret bits extraction. They are based on mitigating error through optimized guard intervals (GI mechanism) or through phase-shifting the channel taps (PS mechanism). The high efficiency of these two mechanisms compared to the regular quantization mechanisms is manifested through simulations based on a realistic channel model.
We also investigate some practical issues that affect the performance of key generation at the physical layer based on the multipath wireless channel. Delay and Mobility are mainly investigated. In fact, mobility leads to a varying channel. Thus, delay between the channel estimation procedures at the two communicating wireless nodes results into varied channel estimates, hence key disagreement. To tackle these two issues, we propose the Enhanced 3-Way PS mechanism. Through simulation results, this mechanism has been proven to be robust to delay and mobility while still achieving a high secret bit extraction rate. Finally, key reconciliation and error correction are also discussed.
The second part of this dissertation is concerned with securing medium access in wireless networks. In fact, the broadcast nature of wireless communications poses a problem with channel access. A selfish node can get easily a higher share of the common wireless channel by simply manipulating through the medium access protocol parameters, mainly the random backoff selection procedure.
To tackle this problem, we first propose the Random Backoff Control (RBC) mechanism. It is based on controlling the backoff selection procedure to ensure a fair distribution of channel resources and enable simple misbehavior detection. The effectiveness of this mechanism in thwarting misbehavior, compared to other related mechanisms, is manifested through simulations based on the OMNeT++ network simulator.
Last but not least, we investigate scheduling-based medium access schemes and we develop the Self-Organized Distributed Channel Access (SODCA) scheme. Intuitively, a scheduling scheme would be resilient to misbehavior and would achieve a higher bandwidth efficiency than contention-based mechanisms. Distinctively from other proposed schemes, SODCA does not incur any additional overhead and is a distributed, efficient, compatible, misbehavior resilient, and a dynamic scheduling scheme. Through simulation results based on the OMNeT++ network simulator, we demonstrate the high efficiency of SODCA compared to contention based mechanisms in both static and dynamic scenarios. | de |