This thesis presents the concept, implementation and evaluation of a cognitive agent that utilizes a mixed-initiative approach for pilot-assisted mission planning. The goal of the agent is to reduce pilot workload, minimize planning errors, and improve situational awareness to ultimately increase mission success rates. The agent is intended to simplify mission pre-planning and support pilots in dynamic (re- )planning tasks. The aim is to make the transition from traditional planning methods in military aviation to an efficient, technology-supported solution. The planning assistance agent acts as a virtual assistant that creates mission plans in real time and makes adjustment recommendations. These proposals are communicated to the pilot through a transparent human-machine cockpit design using cooperative mixed-initiative techniques. Two experiments were conducted with German Air Force pilots in a fighter simulator to test the planning capabilities of the assistant agent and to validate the pilots' interaction in complex military operations. The pilots' preferences in terms of user-friendliness and effectiveness in mission execution were examined at three assistance levels: Management-by-Consent (medium assistance level), Management-by-Exception (high assistance level), and a situation-adapted level where the agent's actions vary according to criticality by either making suggestions or acting directly. The results of the experiments show that the planning assistance agent efficiently creates mission plans that are clear and have minimal errors. The pilots were in favor of using such an agent for mission pre-planning in the future. The planning assistance agent improves mission success and efficiency at all assistance levels, especially with situation-specific assistance. Although the pilots subjectively preferred the high assistance level, the medium assistance level proved to be more successful in terms of mission performance. The planning assistance agent combines human and machine intelligence to significantly improve mission (re)planning for military air operations by increasing pilots' situational awareness and mission success. Despite the positive overall assessment, there is still room for improvement in the creation of plans and presentation of assistance messages, such as more freedom of movement in the target area and adjustments to communication. Overall, the planning assistance agent was very positively received and represents significant progress in dynamic mission planning with cognitive agents.     «

This thesis presents the concept, implementation and evaluation of a cognitive agent that utilizes a mixed-initiative approach for pilot-assisted mission planning. The goal of the agent is to reduce pilot workload, minimize planning errors, and improve situational awareness to ultimately increase mission success rates. The agent is intended to simplify mission pre-planning and support pilots in dynamic (re- )planning tasks. The aim is to make the transition from traditional planning methods in m...     »