A Multi-Task Online Monitoring System for UAV

Due to the limitations of volume, weight and power consumption, the accident rate of unmanned aerial vehicle (UAV) is much higher than that of manned aircraft. Therefore, UAV condition monitoring has become a research hotspot. However, most of the current UAVs do not support the deployment of multiple complex models due to limited airborne resources. In this paper, a multi-task online monitoring system is developed, which can realize real-time analysis of multiple airborne parameters and meet the needs of dynamic monitoring. First, we build different algorithmic models for different monitoring task's requirements, and complete the training of the models. After that, we use High Level Synthesis (HLS) tools to optimize and accelerate each algorithmic model and realize the condition monitoring function. Then, dynamic reconfiguration is adopted to complete the dynamic update of models according to different task requirements. Finally, the dynamically reconfigurable models are deployed in an on-board embedded hardware platform. The experimental results show that the system can realize real-time condition monitoring of multiple tasks.