Implementation of Architecture Concepts for Hardware Agent Systems

Requirements to implement future information processing systems are in particular flexibility, adaptability, and reliability. Agent systems which are known from software (sw) technology are a new way to implement these information processing systems. Features like autonomy, proactiveness, reactivity, artificial intelligence and features for self-organisation of agent systems with focus on self-configuring, self-healing, self-optimizing, and self- protecting make agent systems different from conventional systems. Our goal is to provide architecture concepts for hardware (hw) agent systems, embedded hw platforms to create hw agents with the properties of sw agents, and new methods to make these hw agent systems adaptive and failsafe. This paper describes a specialized embedded hw architecture utilizing for agent systems with a high degree on flexibility, parallelism and reliability.