A Minesweeper Algorithm for Improved Signal Area Estimation in Spectrum Aware Systems

Abstract

Signal area estimation is a critical component of spectrum aware systems. It entails determining the subsets of elements of a time-frequency matrix where a signal is present. This study proposes and assesses the potential of a minesweeper algorithm in estimating accurately the signal area. The proposed method can be employed in two ways: as a standalone signal area estimation technique and also as a pre/post-processing technique in combination with other signal area estimation methods in order to correct signal detection errors before applying the other estimation method (pre-processing) and/or errors introduced by the other estimation method itself (post-processing). The performance of the proposed minesweeper algorithm in both application approaches is evaluated by means of software simulations. The obtained results show that, when used as a standalone method, it can provide similar or even better accuracy than other methods at a much lower computational cost. However, the best performance is obtained when used as a pre/post-processing technique in combination with other existing methods, without increasing significantly the total computation time.