Autonomous vehicle software taxonomy

A survey of software architectures for 12 autonomous underwater vehicles has been made. The analysis is performed in the context of an evaluation framework which focuses on each architecture's important features, including its control distribution strategy, its use of modeling, and the types and representations of knowledge that it uses. A classification scheme has been developed which helps to associate the features of an architecture with vehicle performance characteristics. The manner in which vehicle control and decision-making are distributed among the elements of the software architectures is discussed. A specific onboard planning system architecture is proposed, and its characteristics are discussed in the context of the criteria that were used to evaluate the existing architectures.<
>