The following problem in multiterminal source coding wasintroduced in Berger and Zhang (see 1994 IEEE International Symposium onInformation Theory, Trondheim, Norway). A firm's CEO is interested in adata sequence {X(t)}∞t=1 which cannot beobserved directly. The CEO em ploys a team of L agents who observeindependently corrupted versions of{X(t)}∞t=1. Let R bethe total data rate at which the agents may communicate informationabout their observations to the CEO. The agents are not allowed toconvene. Breger et al. determine the asymptotic behavior of the minimalerror frequency in the limit as L and R tend to infinity. Their resultis for discrete memoryless source and observations. We consider aspecial case of the continuous source and observations problem. Weassume that the source is an i.i.d sequence of zero mean Gaussian randomvariables (𝒩(0,σ2X)) and theobservations are corrupted by identical independent memoryless Gaussiannoise (𝒩(0,σ2N)). The CEO is interestedin reconstructing the source with minimum mean squared error. We studythe asymptotic behavior of the minimum achievable distortion in thelimit as first L and then R tends to infinity
展开▼
