Autonomous Systems Incidentally Controlled by a Remote Operator

Research and development on autonomous systems (AS) focuses the attention of ever more and more countable universities and companies. The autonomy of operation, with respect to the environment and also the complexity of tasks and processes the AS is implemented to operate and solve, requires a knowledge-based approach. However, in many cases, it is very difficult, or even impossible, to equip the system with a complete knowledge base that would allow the flawless operation of AS in any circumstances. Hence, the system can be faced with a situation or a task that it is unable to solve. The approach developed by the author’s team expands the idea of Fault-Tolerant systems. We broaden the meaning of a fault including here such situations, that cannot be solved by the AS on its own. If such a fault is detected, the system assesses its severity, basing on its Knowledge Base, and, once it is unable to solve the problem, contacts the remote operator who takes over the control of the system to overcome the problem. To facilitate remote control of the operator, a Virtual Teleportation technology is applied. Then the control is returned to the AS. This paper deals with the problems of combining autonomy with remote control of the system supported by Virtual Teleportation of the operator. Particular attention is paid to the detection of faults, especially those that are caused by an incomplete Knowledge Base. Additionally, some examples are given.