Locally optimum rank detection of known signals under a weakly dependent noise model is considered in this paper. The weakly dependent noise model is known to be useful for modeling inter-user interference, which is important in the synthesis and analysis of mobile communication systems. For good weak-signal detection performance under any specified noise probability density function, the locally optimum detector is derived based on signs and ranks of the observation. It is also shown that the locally optimum rank detector has the same asymptotic performance as the locally optimum rank detector which uses the actual values of the observations.