Implementation of Threshold Comparator Using Cartesian Genetic Programming on Embryonic Fabric

摘要

Recent research in the area of evolvable design clearly indicates its advantage in the electronics system domain. This biologically inspired approach for design automation and reconfiguration is required for the hardware that needs online adaptation. The Cartesian Genetic Programming (CGP) represents a circuit genotype in the form of grid of nodes. In satellite for safe mode detection, a threshold comparator circuit is used. The comparator circuit can be evolved using CGP architecture by evolutionary algorithm. An evolutionary algorithm (EA) is designed and applied on the CGP pattern of comparator. The evolved comparator in the cascaded form can be further implemented on embryonic fabric. The embryonic fabric has cellular structure that makes it suitable for self-healing and self-replication. The CGP approach to generate configuration data for embryonic array is better than the LUT based data generation approach. In this paper a 2-bit comparator is evolved using customized evolutionary algorithm. The implementation of cascaded 8-bit threshold comparator on the embryonic array is also demonstrated. This comparator design needs configuration data for 2-bit and the data can be cloned to next cells for scalable design on embryonic fabric.