Neighbor discovery for cognitive radio ad hoc networks

Zaw Htike, Choong Seon HongICUIMC 2013 · DOI

TL;DR

This paper proposes a novel, distributed neighbor discovery mechanism for cognitive radio ad hoc networks that adapts its strategy based on the number of available channels. It offers a simpler, more efficient alternative to existing methods by avoiding dedicated control channels or complex channel hopping sequences, leading to a reduced time to rendezvous.

What Problem Does This Paper Address?

Cognitive radio (CR) technology allows secondary users (SUs) to opportunistically use licensed spectrum bands when primary users (PUs) are absent, aiming to improve spectral efficiency. In CR ad hoc networks, SUs need to discover each other to communicate, but this is challenging because the available channels dynamically change due to PU activity. SUs must find a common channel to exchange control information, a process known as neighbor discovery or rendezvous. Existing approaches, like using a dedicated common control channel (CCC) or channel hopping sequences, have drawbacks: CCCs are unreliable due to PU interference, and channel hopping can be complex, require synchronization, and lacks network status awareness. This paper addresses the need for a robust, distributed, and efficient neighbor discovery mechanism that can adapt to the dynamic CR environment without relying on these problematic methods.

Methods

The proposed mechanism is entirely distributed and requires SUs to only have local information, specifically the number of available channels. When an SU needs to discover neighbors, it first performs channel sensing to identify its available channel set. Then, it adopts one of two adaptive strategies:

  • Strategy One: The SU sequentially switches through its available channels, broadcasting a "neighbor discovery message" (DOV) on each free channel and waiting for an acknowledgment (ACK). If no ACK is received, it moves to the next channel.
  • Strategy Two: The SU randomly selects one available channel and waits for a DOV from a neighbor. If a DOV is received, it replies with an ACK.

Crucially, the SU dynamically chooses between these strategies based on the number of available channels it observes. This decision-making process is designed to minimize the expected time to rendezvous (TTR). The paper also analyzes the expected payoffs (in terms of TTR) for different strategy combinations and derives an optimal strategy profile that SUs should adopt.

Datasets

Not discussed in the paper.

Key Insights and Ideas

  • Adaptive Strategy Selection: The core innovation is the dynamic adaptation of neighbor discovery strategies based on the number of available channels. This allows the system to be more efficient in environments with few or many channels.
  • Simplicity and Distribution: The proposed mechanism avoids complex algorithms and centralized coordination, making it highly suitable for ad hoc network environments.
  • No Prior Knowledge Required: Unlike some existing methods, this approach doesn't require SUs to have any prior knowledge of their neighbors' locations or channel hopping sequences.
  • Improved TTR: The adaptive strategy selection and distributed nature lead to a demonstrably lower expected time to rendezvous compared to existing random and sequence-based approaches.

Key Results

The paper presents numerical results comparing the proposed neighbor discovery (ND) mechanism with existing Random Algorithm (RA) and Orthogonal Sequence-based Algorithm (OSA). The results, shown in Figure 4, indicate that the proposed ND mechanism consistently achieves a lower expected time to rendezvous (E[TTR]) across varying numbers of available channels. For instance, as the number of available channels increases, the proposed method's E[TTR] grows more slowly than RA and OSA, demonstrating its superior efficiency.

Limitations and Future Work

The paper acknowledges that the proposed mechanism is designed for scenarios where SUs have no prior knowledge of their neighbors. Future work could explore how to integrate this mechanism with other CR functionalities or investigate its performance in more complex network topologies and under varying PU activity patterns. The paper also implicitly assumes that SUs can accurately sense available channels and that PU activity is not excessively disruptive to the discovery process itself.

Summary updated

BibTeX
@inproceedings{htike2013neighbor,
  author = {Zaw Htike and Choong Seon Hong},
  title = {{Neighbor discovery for cognitive radio ad hoc networks}},
  booktitle = {ICUIMC},
  year = {2013},
  doi = {10.1145/2448556.2448658},
}
Figure