Article Preview
TopIntroduction
Wireless communications play a critical role in our everyday life, e.g., public safety communications. The current public safety communication system is highly channelized. That is, each battalion of first responders form a talkgroup and either a physical or logical channel, depending on whether it is a conventional or trunked radio system, is assigned by the central dispatch center as their main operation channel (Varone, 2003). These channels are regulated by the Federal Communications Commission (FCC). Under normal conditions, these channels may be sparsely occupied. Figure 1 shows an example of public safety channel usage (OKWIN 853.2625MHz and 853.7775MHz) in a typical afternoon in Stillwater, OK. However, during emergent incidents, the assigned channels may become overloaded by talkgroups or the dispatch center may become unavailable due to destructions as evidenced during the U.S. Katrina disaster (Paweczak et al., 2005). It is of great importance that we have a flexible and resilient communication network that can fully utilize the spectrum resources while still be available under critical circumstances.
Figure 1. The spectrumgram of channel usage of the public safety channels at OKWIN 853.2625MHz and 853.7775MHz in a typical afternoon in Stillwater, OK.
Cognitive radios (CR) are smart radios that can dynamically and efficiently utilize available spectrum to communicate. The fixed channels can still be used by the respective services which are known as the primary users. However, during times of low usage, these channels can also be used by cognitive radio users, known as secondary users. Cognitive radios can periodically sense the spectrum to determine which frequencies are available, and then dynamically agree on particular frequencies to communicate on, instead of having every frequency prefixed. A collection of cognitive radios forms a cognitive radio network capable of communicating with each other using the available frequencies.
The main motivation for our study is to improve the communication infrastructure for public safety, which mainly uses a centralized architecture. That is, all communications go through a dispatch center or through the use of repeaters. In an ad-hoc cognitive radio network, all radios are not necessarily directly connected. A routing protocol is then needed for two nodes1 not within the communication range to communicate with each other through other nodes in the network. Compared with existing work, the main challenge lies in that the channel availability in the ad-hoc cognitive radio network is dynamic and uncertain due to primary users' activities.
On one hand, there exist resources including unused spectra and a network of nodes. On the other hand, various communication requests of different priorities and QoS requirements are generated by these nodes from time to time. The important question is how to match the demands and available resources in the most efficient way. In this paper, we first study two main issues for efficient communications: 1) route discovery: identifying multiple possible paths from the source to the destination, and 2) efficient scheduling of spectrum resources on these paths. We adopt the electric field-based routing method to perform route discovery. In multi-hop networks, field-based routing is more preferable as it can avoid congestion, provide load balancing, and more importantly, it is more robust to dynamics of channel availability in a cognitive radio network. Once the destination receives the source request from multiple paths, it replies back with an acknowledgement along each of the paths. Each node along the path also appends its schedule information. The source node will thus have complete knowledge of all the nodes’ schedules for all the paths. The source can then schedule the use of the time-frequency resources such that the transmission delay and the packet loss rate can be minimized. Multiple paths would potentially decrease the transmission time if large amount of data needs to be sent.