Intelligent Event-Triggering Fault Estimation and Fault Tolerant Control of a Quadrotor UAV With Actuator Fault

Quadrotor UAVs have a wide range of applications in production and daily life. To address the issues of observation and fault-tolerant control of UAVs in these applications, we propose a networked control-based event-triggered observation and fault-tolerant control scheme. In our research, by designing dynamic event-triggered thresholds, we can not only observe faults and system states in real time but also implement effective fault-tolerant control compensation based on the observation results. To dynamically optimize the event-triggering thresholds, we use the Proximal Policy Optimization (PPO) algorithm from reinforcement learning for training.By intelligently adjusting the triggering conditions, our scheme not only improves the UAV's robustness to motor faults but also saves computational resources by reducing unnecessary control updates. Through simulations and physical experiments, we have verified the effectiveness of the proposed method. The experimental results prove that the strategy significantly enhances the UAV's fault tolerance capabilities in the event of motor faults, ensuring the reliability and efficiency of UAVs when performing critical tasks. This research provides new technical support for the operation of UAVs in application environments.