Learning-Based Anomaly Detection and Monitoring for Swarm Drone Flights

Abstract

This paper addresses anomaly detection and monitoring for swarm drone flights. While the current practice of swarm flight typically relies on the operator's naked eyes to monitor health of the multiple vehicles, this work proposes a machine learning-based framework to enable detection of abnormal behavior of a large number of flying drones on the fly. The method works in two steps: a sequence of two unsupervised learning procedures reduces the dimensionality of the real flight test data and labels them as normal and abnormal cases; then, a deep neural network classifier with one-dimensional convolution layers followed by fully connected multi-layer perceptron extracts the associated features and distinguishes the anomaly from normal conditions. The proposed anomaly detection scheme is validated on the real flight test data, highlighting its capability of online implementation.

A Survey on Fault Diagnosis and Fault-Tolerant Control Methods for Unmanned Aerial Vehicles

“…The work in [139] discusses anomaly detection and monitoring on swarm drone flights and provides a machine-based learning framework to detect abnormal conduct of a wide range of flying drones. The approach operates in two stages and the anomaly detection system was validated on actual flight test data, while its ability to run online has been emphasized.…”

Deep-Learning-Based Approach to Anomaly Detection Techniques for Large Acoustic Data in Machine Operation

“…Glowacz et al [ 13 ] classified acoustic feature vectors with the nearest neighbor classifier and naïve Bayes classifier. Another prior study [ 11 ] presented a method for labeling test data obtained through multiple flight tests as normal and abnormal using the K-means method to generate balanced learning datasets and then used that dataset to classify states in real time based on the linear regression method. However, in a real environment, the anomaly situation does not appear more often than the normal situation; thus, it is difficult to collect sufficient data on the anomaly state.…”

“…However, determining the inspection index in the process of automating the noise and vibration quality inspection is not easy and requires a lot of time and effort. For this reason, in recent years, deep learning models have been used to minimize the complexity of data preprocessing, feature extraction, and feature selection [ 11 , 12 ].…”

Special Issue on Unmanned Aerial Vehicles (UAVs)

“…W. Yue, X. Guan, and L. Wang present the UAV cooperative search mission for multi-dynamic targets in sea areas using a reinforcement learning (RL) algorithm [21]. H. Ahn, H. Choi, M. Kang, and S. Moon present a machine learning-based framework to detect and monitor abnormal behaviors during UAV swarm operation [22].…”