Towards the design of efficient and versatile cognitive robotic architecture based on distributed, low-latency working memory

Autonomous robots will be present in our daily lives in the coming years. One of the most critical elements facilitating this expansion of robots is the concept of Cognitive Robotic Architectures (CRAs). Thanks to these CRAs, robots are aware of their state and surroundings and then build all the behaviors best suited to the scenario on this information. In recent years our team has proposed an CRA called CORTEX designed for use with autonomous robots working in human environments. CORTEX is based on a distributed graph-like working memory where software agents can read or update information. In this paper, we describe the design process of the new CORTEX architecture up to its current implementation. Among the most salient design requirements is data synchronization between the different agents in the architecture, low latency, and performance of the new architecture. To validate the effectiveness of the architecture and its versatility, we have used our CRA in different use cases, including social robot navigation and autonomous driving of connected vehicles.