Low-power state assignment for asynchronous finite state machines

In this paper a precise model for calculating the steady state probabilities and state transition probabilities of asynchronous finite state machines (AFSMs) is presented. Based on this model and the derived weighted adjacency diagram, an efficient unicode race-free state assignment technique is proposed to minimize the average switching activity of state variables, thus to potentially reduce the average power dissipation in AFSMs. Experimental results show that the proposed algorithm can achieve less switching activity of state variables in comparison with existing unicode state assignment techniques for AFSMs.