Embryonics: multi-cellular and multi-molecular digital systems

Drawing inspiration from biological systems for the design of robust digital circuits is far from being a novel concept: in the 1950s, J. von Neumann (1966) began investigating the biological concept of self-replication in order to design robust machines. The design of machines capable of executing very complex tasks and at the same time capable of self-replication, however, proved to be too complex for the technology of the 1950s, and von Neumann´s project never went beyond a theoretical study. The Embryonics project (G. Tempesti, 1998; G. Tempesti et al., 1998) is an attempt at realizing machines which fulfil von Neumann´s ambition to design robust self-replicating machines, but exploit a different approach so as to become feasible given today´s technology. The main feature of our Embryonic systems is the idea of structuring a digital machine (a digital circuit dedicated to the execution of an arbitrarily complex task, equivalent to a biological organism) as an array of identical processors (the digital equivalent of a biological cell), capable of altering their size and structure to fit the requirements of the task at hand. This adaptability is achieved by implementing the cells themselves using an array of small, programmable digital elements (a field-programmable gate array, or FPGA, whose elements are analogous to biological molecules)