Throughput Maximization with Coverage Bounds for Lifetime-Constrained Multi-Channel Cooperative Spectrum Sensing Scheduling

Abstract

Cooperative spectrum sensing scheduling has been studied mainly as a lifetime-maximization problem. Many monitoring deployments, however, face the converse requirement: the operating horizon is fixed in advance, and the question is how much spectrum opportunity the network can deliver across many channels while lasting that long. This paper formulates that converse problem, with the binary assignment of secondary users to channels over time as the decision variable, aggregate throughput as the objective, and the required horizon, per-node energy budgets and per-channel detection and false-alarm levels as constraints. Cooperation bounds are derived for a general k-out-of-N fusion rule by a stochastic-dominance argument; they hold for heterogeneous detection quality and reduce to the known OR-rule results at k = 1. These bounds yield a guarantee and a converse on the number of channels that can be monitored for the required horizon, together with an exact flow-based certificate of achievability that meets the converse in the studied regime. Simulations show that a time-shared sensing schedule improves throughput by 29.7% (95% confidence interval - CI [28.6%, 30.8%]) over a lifetime-aware static assignment at the design horizon of 9.72 h, in all 200 paired realizations. Because a repaired lifetime-agnostic assignment performs identically to the static one, the gain appears to stem from time sharing rather than from lifetime awareness. The ratio of time-shared to static energy supply predicts the gain to within 4.7% without fitted parameters, and the schedule attains 98.9% of the proved converse on average.