Abstract:The sensory system of unmanned aerial vehicles (UAVs) plays an important role in flight safety. Thus, any sensor fault/failure can have potentially catastrophic effects on vehicle control. The recent advance in adversarial studies demonstrated successful sensing fault generation by targeting the physical vulnerabilities of the sensors. It poses new security challenges for sensor fault detection and isolation (FDI) and fault recovery (FR) research because the conventional redundancy-based fault-tolerant design is not effective against such faults. To address these challenges, we present a redundancy-free method for UAV sensor FDI and FR. In the FDI design, we used a basic state estimator for a rough early warning of faults. We then refine the design by considering the unmeasurable actuator state and modeling uncertainties. Under such novel strategies, the proposed method achieves fine-grained fault isolation. Based on this method, we further designed a redundancy-free FR method by using complementary sensor estimations. In particular, position and attitude feedback can provide backup feedback for each other through geometric correlation. The effectiveness of our approach is validated through simulation of several challenging sensor failure scenarios. The recovery performance is experimentally demonstrated by a challenging flight tasks-restoring control after completely losing attitude sensory feedback. With the protection of FDI and FR, flight safety is ensured. This UAV security enhancement method is promising to be generalized for other types of vehicles and can serve as a compensation to other fault-tolerant methodologies.

Asymmetric Dual Redundant Inertial Sensors: A Method for F Ault Detection and Diagnosis

A Novel Fault Detection and Isolation Technology for Redundant Inertial Measurement Unit

Real-Time Attitude Sensor Fault Detection Based on Nearest Neighbor Reference Set and E2LSH Acceleration

Fault Detection and Isolation for Redundant Inertial Navigation System Based on Asymmetric Precision

Research on fault detection for redundant inertial measuring unit based on the generalized likelihood ratio

Application of Analytic Redundancy-based Fault Diagnosis of Sensors to Onboard Maintenance System

Innovative Soft Fault Diagnosis Method for Dual-Redundancy Sensors

Fault-Tolerant SINS/Doppler Radar/Odometer Integrated Navigation Method Based on Two-Stage Fault Detection Structure

Fault Detection of Redundant SIMU Based on DBN

Fail Self-Diagnosing Technique In Closed-Loop I-Fog And It'S Application In Redundant Strapdown Inertial Measure Unit

A Method for Comprehensive Diagnosis of Sensors of Flight Control System Using Analytical Redundancy

ALPF-GLRT Based Fault Detection Method for Small Faults Applied to Redundant IMUs

An IMU Fault Diagnosis and Information Reconstruction Method Based on Analytical Redundancy for Autonomous Underwater Vehicle

Sensor failure detection and isolation for control system of underwater vehicle

Failure Detection of Redundant Sensor with Reduced-order Parity Vectors

A Diagnosis Algorithm for Abnormal Data of Spin-Stabilized Satellite Attitude Sensors

Innovation-based Kalman filter fault detection and exclusion method against all-source faults for tightly coupled GNSS/INS/Vision integration

ARAIM Based on Fault Detector Reuse for Reducing Computational Load

Redundancy-Free UAV Sensor Fault Isolation and Recovery.

Multiple Sets of INS Fault Detection Based on Information Fusion

Orthogonal Experiment of RIMU Fault Detection