In a proposed method of designing a recon- figurable antenna, the design would be opt- imized through evolution in hardware. The proposed method would be a specific instance of an emerging general method of automated synthesis of electronic circuits in hardware. Other specific instances of the general method were described in two NASA Tech Briefs articles: “Reconfigurable Arrays of Transistors for Evolvable Hardware” (NPO-20078), Vol. 25, No. 2 (February 2001), page 36; and “Evolutionary Automated Synthesis of Electronic Circuits” (NPO-20535), which precedes this article. To recapitulate: Under the direction of genetic and/or other evolutionary algorithms, the configurations and thus the functionalities of circuits would be made to evolve until at least one circuit exhibited a desired behavior. Evolution would include selective, repetitive connection and/or disconnection of transistors, amplifiers, inverters, and/or other circuit building blocks.
According to the proposed method, a reconfigurable antenna in a basic initial configuration would be placed on an antenna test range equipped for testing at the frequency or frequencies of interest. A computer outside the test range would be connected to interface circuits that would, in turn, be connected to (1) the test equipment (transmitters and receivers) on the range and (2) wires through which the computer could control the configuration of the antenna.
The computer would execute software that would include one or more automated- optimization algorithms plus driver-interface software for controlling the antenna configuration and the test equipment. Following initial activation, the software would go through the optimization process, controlling the test equipment and the antenna configuration as needed to produce an optimized configuration for each set of desired electromagnetic properties.
The only other method of automated design by use of an optimization algorithm involves computational simulation of performance instead of testing of a real physical implementation. The proposed method does not involve computational simulation and is expected to surpass the method that involves computational simulation; this is because the results of testing a real physical implementation are inherently valid and more accurate than are results obtained through computational simulation. In addition, optimization by use of the proposed method is expected to take much less time than does optimization by use of a computational simulation of reasonable fidelity. Moreover, unlike in the computational-simulation method, there would be no need to try to validate the simulated results with a physical test — an undertaking that usually entails manual re-optimization of the design to obtain the same performance from the physical device as from the simulated one.
This work was done by Adrian Stoica and Derek Linden of Caltech for NASA’s Jet Propulsion Laboratory.
This Brief includes a Technical Support Package (TSP).
Designing Reconfigurable Antennas Through Hardware Evolution
(reference NPO-20666) is currently available for download from the TSP library.
Don't have an account? Sign up here.