Development and Testing of a Complementary Sensor Network for Robust Estimation of Maneuver and Gust Loads

Abstract

In this publication, a sensor and observer network is presented with the primary objective of increasing the robustness of structural loads estimation. This augmentation is achieved through the combination of measurement methodologies and model-based load observers, thereby creating synergistic effects that mitigate the limitations associated with each approach. This work outlines the development of a complementary sensor network, comprising laboratory tests and virtual flight tests. The sensor technologies employed include strain gauges, fiber bragg sensors, inertial measurement units, camera-based optical deformation measurement, and MEMS pressure measurement profiles. For each of these technologies, the laboratory test development and testing process, alongside the derivation of sensor models for virtual testing of the sensor network is presented. Within the context of the sensor network, these redundant and partially complementary sensors are fused through the utilization of both local and central Kalman filters. The local fusion strategy exploits the integral correlation between inertial measurement unit (IMU) and camera data at corresponding observation points, establishing the basis for employing a data-driven local-model network approach wherein local deformations are trained on structural loads data. The central load fusion combines a data association algorithm based on a quadruple-voting scheme and an extended Kalman filter. Based on virtual flight tests considering a load sensor failure, the performance and robustness of the whole sensor network is demonstrated.