3-DoF Longitudinal Aircraft System Identification

A MATLAB implementation of a classical system identification pipeline for estimating longitudinal aerodynamic parameters of a fixed-wing aircraft from flight data. This serves as a baseline reference implementation. The results it produces are intended to provide a basis of comparison for other system identification methods.

The workflow uses two sequential stages: an Equation-Error method for initial parameter estimates, followed by an Output-Error method for maximum-likelihood refinement.

---

Files

File	Description
AC_VER2_main.m	Script to call the other files
AC_VER2_sysID.m	Main sysID implementation — runs both equation-error and output-error methods
AC_VER2_statespace.m	Supporting file defining the nonlinear 3-DoF fixed-wing state-space model
AC_VER2_statespace.m	Supporting file for plotting states and inputs
model.py / classical_OEM_resultsforAIAA 1	Files used to generate the flight data

---

Pipeline

Flight Data
    │
    ▼
Equation-Error Method  —  constrained least squares (lsqlin)
    │                     decomposes into lift, drag, and moment sub-problems
    ▼  initial parameter estimates
Output-Error Method  —  oe.m (SIDPAC)
    │                   Modified Newton-Raphson + Simplex fallback
    │                   Full nonlinear RK4 simulation at each iteration
    ▼
Final Parameter Estimates + Cramér-Rao Bounds

---

Parameters Identified

CL0, CL_alpha               — lift model
CD0, k_CD                   — drag polar
CM0, CM_alpha, CM_Q, CM_e   — pitching moment model

---

Dependencies

• MATLAB + Optimization Toolbox (lsqlin)
• SIDPAC toolbox (oe.m and supporting routines: mnr.m, simplex.m, estrr.m, misvd.m, cvec.m, compcost.m)
• deriv.m — smoothed numerical differentiation
• lesq.m — unconstrained least squares