An evolutionary approach to design and calibration of reconfigurable OTA-C filters

Latest trends in communication and consumer electronics target the development of fully adaptive systems on mixed signal systems-on-chip. This paper proposes an evolutionary approach to adaptive filter design. The hardware support of the evolutionary design algorithm is a field programmable analog array (FPAA). The FPAA has been implemented with OTA-C techniques. The synthesis algorithm has been developed considering the particularities of the proposed FPAA and of the supported filter structures. Topological reconfiguration is performed with a lookup table, parametric design is performed with evolutionary computation methods and wide-band self-calibration is achieved by means of a phase locked loop. The filter sizing algorithm is suitable for both intrinsic or extrinsic evolution and has the potential to be hard-wired in dedicated digital circuitry. Extensive transistor level simulations performed on the evolved adaptive filters prove the functionality of the proposed software and the underlying circuitry.