Simulation data and code for platform-coupled kinematic model predictive control of a shipborne two-axis antenna gimbal

Description

This dataset contains simulation data, Python and MATLAB source code, Simscape Multibody models, numerical tables, and figures supporting the manuscript “Platform-Coupled Kinematic Model Predictive Control of a Shipborne Two-Axis Antenna Gimbal under Vessel Motion”. The study evaluates horizon-wise inverse-kinematic reference generation under prescribed vessel roll, pitch, yaw, and heave. The data cover nominal controller comparisons, preview and disturbance-feedforward ablations, prediction-horizon and computational-delay sweeps, model mismatch, encoder noise, actuator constraints, extended simulation records, and 100 paired Monte Carlo trials. The nominal Python controller achieves an RMS geodesic pointing error of 0.517393 mrad. The independent Simscape Multibody record yields 0.575863 mrad after common-grid reduction. All records are numerical simulations; no field measurements or hardware experiments are included. Synthetic motion-severity labels do not represent vessel-specific, hydrodynamically validated sea states. During the September 2026 review, the Python campaigns were replayed, including all 100 paired Monte Carlo trials. Saved MATLAB/Simscape trajectories were independently reprocessed; the MATLAB/Simscape dynamic models were not rerun. Superseded exact-preview timing results are identified in the README. The archive includes a data dictionary, software requirements, a portable verification script, figure-generation code, and a SHA-256 file manifest. See README.md for reproduction instructions and methodological limitations.

Steps to reproduce

1. Extract PCK_MPC_Data_and_Code_v1.zip and open the extracted dataset directory. 2. Read README.md and DATA_DICTIONARY.md for the directory structure, variable definitions, units, software requirements, and methodological limitations. 3. Install the Python dependencies listed in requirements.txt. The versions used during verification are recorded in requirements-tested.txt and runtime_versions.json. 4. Run: python verify_dataset.py This checks the stored numerical results against the reference values included in the dataset. 5. To replay the six nominal Python controller configurations, run: python verify_dataset.py --nominal 6. Follow the campaign entry points documented in README.md to reproduce the Monte Carlo trials and parameter sweeps. Use the supplied plotting scripts to regenerate figures. 7. MATLAB/Simscape source files, models, and saved trajectories are included. Re-running these models requires the relevant MATLAB/Simulink/Simscape products. During the September 2026 review, saved trajectories were independently reprocessed, but the dynamic MATLAB/Simscape models were not rerun. All results are numerical simulations, not experimental or field measurements. Historical diagnostic results and superseded timing configurations are identified in README.md.

Institutions

• İskenderun Technical University

  Hatay, İskenderun

Categories

Licence