Spectrum sharing with an incumbent provider imposes constraints on entrants to limit the associated interference. Existing designs use conservative models of the propagation environment and worst-case entrant power levels for ensuring protection, leading to spectral inefficiency. In this paper, we present a new approach for estimating entrant transmission powers based on real-time distributed sensor measurements. This knowledge is vital information that can then be used to infer aggregate interference at an incumbent receiver location, or for other purposes such as such as spectrum monitoring and enforcement. We propose a computationally efficient two stage estimation algorithm based on a maximum likelihood objective with log-normal fading and exponential noise, and present numerical results illustrating performance.
