In this paper

the well known signal subspace projection concept proposed by Schmidt is introduced to analyse the array resolution capacity or the maximum number of signals resovable by a given uniform linear array as well as the minimum number of sensors of a uniform linear array for resolving a given group of plane waves. Special attentions are paid to determination of signal subspace in coherent signal environments. We have shown that

although the signal subspace in N （the number of sensors） dimensional vector space can not be determined in the coherent signal environment

the signal subspace in lower dimentional vector space could be constructed

and this can be carried out in three ways. Based on the above concept

the limitations on the two array resolution parameters are derived in the paper. The derivation is straight forward and intuitive. Examples are also given for illustration.