Intelligent Control Architecture for Autonomous Vehicles
The use of remotely-operated vehicles is ultimately limited by economic support costs, and the presence and skills from human operators (pilots). Unmanned craft have the potential to operate with greatly reduced overhead costs and level of operator intervention. The challenging design is for a system that deploys a team of Unmanned Vehicles (UVs) and can perform complex tasks reliably and with minimal (remote) pilot intervention. A critical issue to achieve this is to develop a system with the ability to deal with internal faults, and changes in the environment as well as their impact on sensor outputs used for the planning phase.
The tutorial objective is to present step by step the development process (from requirements to prototyping) of an Intelligent Vehicle Control Architecture (IVCA) that enables multiple collaborating UVs to autonomously carry out missions. The architectural foundation to achieve the IVCA lays on the flexibility of service-oriented computing and agent software technology. An ontological database captures the remote pilot skills, platform capabilities and, changes in the environment. The information captured (stored as knowledge) enables reasoning agents to plan missions based on the current situation. The combination of the two above paradigms makes it possible to develop an IVCA that is able to dynamically reconfigure and adapt itself in order to deal with changes in the operation environment. The ability to perform on-the-fly re-planning of activities when needed increases the chance to succeed in a given mission. The IVCA realization is underpinned by the development of fault-tolerant planning and spooling modules (fault diagnosis and recovery) as well as a module called matchmaker to link services with available capabilities.
The IVCA is generic in nature and can be easily adapted to UVs from different domains (i.e. land, water, and air/space). However, the IVCA aims at a case study where Unmanned Marine Vehicles (UMVs) are required to work cooperatively. They are capable of cooperating autonomously towards the execution of complex activities since they have different but complementary capabilities. The above UMV configuration, where the marine robots are tasked to autonomously do mission works before recovery, is possible at a cost of endowing the UMVs with “intelligence” that in former solutions is provided by remote or even in-situ human pilots.
The IVCA development applies the software/systems engineering principles. The tutorial is structured in four parts. Part I (background) consists of a brief review of technologies related to the IVCA and a comparison of control architectures for autonomous UVs. Part II (requirements analysis and design) entails the user and system requirements, and the system architecture specification/design. Part III (implementation and integration) describes the IVCA realization based on Robot Operating System (ROS) for the above case study. Session IV (verification and validation) deals with the evaluation of the IVCA by means a simulation.