User Interaction with Intelligent UCAV Agents

摘要

Semi-autonomous Unmanned Combat Air Vehicles (UCAVs) have the potential to deliver increased mission capability by allowing a single operator to supervise multiple UCAVs. However, it is impossible to assume that a single UCAV, let alone several, can be controlled in minute detail. Perhaps the most effective way to use UCAVs in an operational context is to support the operator with high levels of automation and rely upon intelligent software that sits onboard the UCAVs. QinetiQ's research approach has been to concentrate initially on the viability of controlling multiple platforms from a single operator station, identifying the level of command abstraction and information feedback required to enable operators to manage complex missions effectively and with a reasonable level of workload. Such viability has been established but it is clear that the operator will need to have a high level of confidence in the automation if workload is to remain manageable. The next stage of the programme has involved development of advanced software techniques, including cooperative agent and dynamic resource allocation techniques, to provide the necessary onboard intelligence. This brings a danger that the operator will be unaware of what the UCAVs are doing and/or planning at particular stages of the mission and, consequently, not feel in full control. This could lead to reduced mission effectiveness and potentially to UCAV loss. The challenge is to provide information in a way that gives the operator greater confidence, by providing visibility of UCAV behaviour and intent, without increasing workload unduly.