Research on Optimization Generation Methods for Test Scenarios of UAV Mission Replanning Capability

As mission environments grow increasingly complex and uncertain, there is a pressing need for comprehensive, systematic, and scientific testing and evaluation of unmanned aerial vehicle (UAV) intelligence. Task re-planning capability, a key metric for assessing UAV intelligence, relies heavily on the design of test scenarios, which directly impacts the accuracy and efficiency of evaluation outcomes. This paper proposes a test scenario generation method for UAV task re-planning capability based on proximal optimization algorithms. By integrating rule constraints and a model-driven approach, the method enables efficient generation and optimization of test scenarios. The introduction of rules and network models allows for the rapid construction of diverse and complex mission environments, significantly enhancing testing efficiency. Experimental results demonstrate that the proposed method excels in scenario diversity and test coverage, providing robust technical support for the evaluation of UAV task re-planning capabilities.