Research on autonomous solution methods for unmanned autonomous systems considering performance constraints

Abstract

In order to elucidate the constraints of autonomy, this article starts from the definition of boundaries and speculates on the autonomy of unmanned systems based on performance constraints. Furthermore, it combines with the OODA loop to explain the hardware and software constraints of autonomy throughout the entire process of perception, planning, decision-making, and control from a system perspective. Finally, through three autonomy tasks, including control allocation, flight control, and path planning, it specifically illustrates the impact of basic capabilities of unmanned platforms, artificial intelligence capabilities, and safety constraints on autonomy. The results indicate that autonomy can only be effective within certain constraints, and once these constraints are exceeded, autonomy becomes difficult to achieve. This suggests that when studying autonomy problems, the practical significance of autonomy should be considered based on the performance constraints of the system, and optimization methods should be employed to find solutions to access the feasibility of autonomy.