基于HMM和DT的无人机异常检测方法*

为了实时检测无人机异常状态，提出基于隐马尔可夫模型(Hidden Markov Model,HMM)和决策树(Decision Tree,DT)的无人机异常检测方法(HMMDT)。首先根据异常致因将无人机异常分为干扰异常和硬件异常；然后结合HMM和DT建立无人机异常检测模型，定义无人机异常度衡量异常状态的严重程度，确定其阈值作为异常分类标准；最后用经纬600pro型无人机进行实操验证，该方法异常检测召回率达92.9%，准确率达97.2%;对硬件异常的识别准确率达88.2%。结果表明:与传统异常检测方法相比,该方法在可以满足无人机实时异常检测需要的同时，具有较高的检测准确率和较小的时间复杂度。

Anomaly detection method of UAV based on hidden Markov model and decision tree

In order to detect the UAV anomaly in real time,a UAV anomaly detection method (HMMDT) based on the hidden Markov model (HMM) and decision tree (DT) was proposed.Firstly,the UAV anomaly was divided into interference anomaly and hardware anomaly according to the anomaly causes.Secondly,the UAV anomaly detection model was established by combining HMM and DT,Then the UAV anomaly degree was defined to measure the severity of anomaly state,and its threshold was determined as the anomaly classification standard.Finally,the Jingwei 600Pro UAV was used to verify the method.The recall rate and precision of anomaly detection by the method reached 92.9% and 97.2% respectively,and the recognition accuracy of hardware anomaly reached 88.2%.The results showed that compared with the traditional anomaly detection methods,this method could meet the needs of UAV anomaly detection in real time with higher detection accuracy and smaller time complexity.