The primary application of autonomous radio for deep space network is relaying communication signals from multiple deep space assets. The advantage of an autonomous radio is that it can communicate to each asset that comes into view, automatically, without having to be reconfigured from Earth for each pass to account for differences in the signal characteristics. The autonomous radio that can meet the need for deep space network is discussed. The fundamental difference between a conventional radio, or even a software-defined radio, and a truly autonomous radio for deep space network is that an autonomous radio has the ability to recognize features of an incoming signal without much a priori knowledge. Using conventional estimation and tracking design, one quickly gets into a chicken and egg problem. In this paper, we overview the architecture of an autonomous radio, define many parameters one might desire to estimate from the signal. Finally, an iterative layered architecture model of parameters estimating in the autonomous radio is presented, message-passing and the interaction of the estimators is analyzed.