State Space Reconfigurability: A Low Energy Implementation Architecture for Self Modifying Finite Automata

Many embedded systems contain state-space disjoint behavior slices with mutually exclusive schedule. When such behaviors are captured by state machines, the current design flow will capture it as a union of all the behavior slices. A traditional state assignment followed by logic synthesis on disjoint behavior slice state sets results in high area and energy costs. Such implementation's costs are proportional to the union of all the behavior slices (in area, energy and delay). We propose to use self-modifying finite automata (SMFA), that have been studied from complexity-theoretic perspective, for expressing and implementing such adaptive behaviors in embedded systems. Towards this end, we present an implementation architecture for SMFAs. We present five adaptive behaviors captured by SMFA to illustrate the expressivity of this formalism. We also compare the area, time and energy costs of several SMFA implementations with the classical logic space (FSM) implementations for some randomly generated state machines. The key contributions of the paper are in exploring the expressibility of self-modifying FA paradigm in realistic embedded systems, and in proposing and evaluating an implementation architecture for SMFAs.