Learning autonomy: Advanced autonomy systems must have self-learning capabilities and be able to improve control performance through autonomous correction, optimization, and learning behaviors based on objects, environments, tasks, and control effects.
Therefore, advanced autonomous systems have the characteristics of adaptability, self-repair, intelligence, collaboration, self-learning, etc.
Autonomous control includes automatically completing pre-determined routes and planned tasks, or online perception, and making decisions in flight and autonomously performing tasks according to the determined missions and principles. The challenge of autonomous control is to solve a series of optimal solution problems in real time or near real time under the condition of uncertainty, and no human intervention is required. The automatic decision in the face of uncertainty is a logical level of progress in autonomous control from internal loop control, autopilot to flight management, multi-aircraft management, and then mission management. It is also the automatic control from the continuous response control level to discrete An extension of the event-driven level.
