Abstract
To account for the randomness of propagation channels and interference levels in hierarchical spectrum sharing, a novel approach to multihop routing is introduced for cognitive random access networks, whereby packets are randomly routed according to outage probabilities. Leveraging channel and interference level statistics, the resultant cross-layer optimization framework provides optimal routes, transmission probabilities, and transmit-powers, thus enabling cognizant adaptation of routing, medium access, and physical layer parameters to the propagation environment. The associated optimization problem is non-convex, and hence hard to solve in general. Nevertheless, a successive convex approximation approach is adopted to efficiently find a Karush-Kuhn-Tucker solution. Augmented Lagrangian and primal decomposition methods are employed to develop a distributed algorithm, which also lends itself to online implementation. Enticingly, the fresh look advocated here permeates benefits also to conventional multihop wireless networks in the presence of channel uncertainty.
Introduction
In Cognitive Radio (CR) networks, unlicensed secondary users (SUs) must communicate without disrupting licensed Primary Users (PUs). This task is complicated by the inherent unpredictability of wireless channels due to fading and shadowing. In hierarchical setups, SUs often lack perfect information regarding PU activity. Traditional routing protocols, which often rely on deterministic "disk models" or fixed link capacities, frequently fail in these dynamic environments where links can vanish instantly due to PU interference or shadowing.
This paper addresses the challenge of "spectrum-cognizant" routing by treating channel randomness and interference as statistical variables. The goal is to jointly optimize routing, medium access control (MAC), and transmit power to ensure reliable data delivery while strictly protecting PUs.
Statistical Routing Methodology
The proposed framework moves away from deterministic routing toward a Statistical Routing approach. Instead of a single fixed path, packets are forwarded based on probabilities that account for the statistical likelihood of link availability.
System Model and Flow Conservation
The core of the optimization framework relies on maintaining queue stability at each node. As shown in Figure 1, for any node $n$, the sum of exogenous traffic $\rho_n$ and incoming traffic from neighbors $j$ must be balanced by the outgoing traffic to neighbors $i$.
The stability constraint ensures that the average arrival rate does not exceed the average service rate, which is a function of the transmission probability $\lambda_n$, routing probabilities $r_{n \to i}$, and the link success probabilities $t_{n \to i}$.
Cross-Layer Optimization
The framework jointly optimizes three layers:
- Routing: Determining the probability $r_{n \to i}$ of forwarding a packet to a specific neighbor.
- MAC: Optimizing the probability $\lambda_n$ that a node attempts to transmit in a given time slot.
- Physical Layer: Calculating the optimal transmit power to manage interference and ensure link reliability.
To handle the non-convexity of the resulting optimization problem, the authors employ Successive Convex Approximation (SCA). This method transforms the complex problem into a sequence of simpler convex subproblems. Furthermore, to enable scalability and online implementation, the Alternating Direction Method of Multipliers (ADMoM) and primal decomposition are used to create a distributed algorithm where nodes only need to exchange local information with immediate neighbors.
Performance Evaluation
The framework was evaluated using simulations of a 7-node secondary network operating in the presence of two active PU transmitters. The environment included path loss (exponent 3.5), log-normal shadowing (6 dB standard deviation), and Rayleigh fading.
Key Results
- Outage-Awareness: The algorithm successfully identifies reliable paths rather than just the shortest ones. It naturally avoids nodes near PUs where interference constraints would force low transmit power and high outage probabilities.
- Convergence and Efficiency: The distributed ADMoM algorithm typically reached consensus within 8 to 15 iterations. This statistical approach allowed for higher exogenous traffic rates compared to protocols that ignore channel randomness.
- Interference Protection: Primal decomposition effectively enforced PU interference power constraints (e.g., -80 dBW), ensuring the secondary network remained "invisible" to licensed users.
Conclusion
This research provides a robust mathematical foundation for multihop routing in uncertain wireless environments. By leveraging outage probabilities and statistical modeling, the proposed framework ensures reliable communication in cognitive radio networks. While currently optimized for "slow" environmental dynamics, future work aims to extend these principles to highly mobile ad-hoc networks (MANETs) where topological changes are more rapid.
Summary updated