Post-flight trajectory, wind and aerodynamic angles reconstruction for launch vehicles in atmospheric flight

The research deals with the topics of trajectory reconstruction (TR) and post-flight estimation of aerodynamic angles and wind components for launch vehicles (LV) during the atmospheric phase of the ascending flight. The aim is to investigate stateof-the-art methods used in TR and post-flight analysis, assessing their limitations to improve existing approaches and developing suitable and effective algorithms. A complete nonlinear simulation model of the LV motion, featuring a realistic control system, comprehensive of all relevant aspects for the post-flight analysis is first developed to generate synthetic sensor data to be used for testing and validation of the proposed estimation methodologies. The simulation environment is completed with models of the atmosphere, wind, and sensors Among the algorithms for TR referenced in the literature, one of the most widely used for TR is NewSTEP. Starting from the NewSTEP kinematic formulation for the process model, an Extended Kalman Filter (EKF) and an Unscented Kalman Filter (UKF) featuring a Fraser-Potter smoother are developed to achieve the best-estimated trajectory (BET). The filters make use of vehicle and ground-based measurements, where acceleration and angular rates measured by an Inertial Measurement Unit (IMU) are treated as inputs in the filter model; uncertainty parameters such as IMU biases are also included as filter states. The consistency and performance of smoothed-UKF and EKF estimates are evaluated and compared by means of Monte Carlo analyses for the case study involving the first-stage flight of the Ares-I LV, in a mission to the International Space Station (ISS) starting from the Kennedy Space Centre Complex LC-39B. The estimation of aerodynamic angles and wind components is realized by exploiting a technique based on a Consider Extended Kalman filter (CEKF). The filter uses measurements provided by an IMU, a GNSS sensor, and an Air Data System (ADS), while first-order Gauss-Markov processes allow for the evaluation of elastic deformations at the gimbal and the wind modeling. The formulation takes into account the systematic errors of the IMU as biases and scale factors (selected as parameters for the CEKF), whereas the airspeed components are considered as filter states in place of the aerodynamics angles, which are computed a-posteriori. A multiplicative error quaternion formulation is adopted to provide an accurate yet robust reconstruction of the LV attitude. The filter performance for the Consider Multiplicative Extended Kalman filter (CMEKF) is assessed by specific simulations and Monte Carlo campaigns. Results are presented and discussed for the aforementioned phase of the mission of the reference LV model. The methodology appears suitable for the estimation of the angle of attack, sideslip angle, and wind speed components, in spite of high levels of uncertainty on model parameters and wind data.