Interactive Immunity and Adaptive Innovation

Autonomous systems must have the ability to operate on their own in dynamic, uncertain environments without breaking down. This paper presents concepts for the design and evaluation of selfrepairing systems that draw on current work on immune mechanisms and artificial immune systems. To survive in a dynamic environment we say that a constructed system must be able to adapt. The problem is how to implement the adaptive drive. This paper looks into one possible route: to model and implement features inspired by the immune system as a problem solving mechanism. This route does not exclude other adaptive mechanisms and can complement them. The autonomous system detects malfunctions and tries to fix them on its own, tinkering with all it has at hand. Internal innovation happens when a new type of malfunction is fixed. An interesting aspect of this process is that what causes a new type of malfunction can be the result of the system’s interaction with a new environment. Therefore this immune response mechanism functions as an adaptive drive. The system remembers the solution for future use and with quicker response. It has innovated with respect to its previous capabilities, and learns from this action. In this sense the system has adapted.