Delayed crowd-voted control with a contrarian minority: code, protocols, simulation outputs and hardware logs

Code, protocols, simulation outputs and hardware logs for the article "Delayed crowd-voted control with a contrarian minority: a lead-compensator model and its transfer from simulation to physical robots" (BongKeun Song, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU)). When many participants steer one robot by voting on its heading from a delayed view, the aggregate lags and tracking degrades. The article shows that a minority voting against the previously published aggregate turns the aggregation into a first-order lead compensator, gives its gain, lead time and stability limit in closed form, and tests the model in more than 140,000 runs on a vote-level engine, the Robotarium simulator and a MuJoCo robot model, with three crowd models and four host controllers (two from Nav2), and on eight physical robots of the Robotarium testbed. Analysis rules were fixed before every run. Contents:- P2prime_code_and_simulation.zip (317 files): the simulation engine, crowd models, runners, protocols with their SHA-256 freeze records, raw results (.jsonl, .npy), analysis scripts and verdicts of every simulation experiment, the ROS 2 Nav2 package, and the figure scripts and figures.- P2prime_hardware_robotarium.zip (309 files): the frozen hardware analysis plan and records, the Robotarium upload scripts, the hardware logs, the hardware-faithful simulation mirror and the agreement analysis.- P2prime_hardware_offline_sims.zip (311 files): pilot dry runs, gap checks and the offline mirror calibration.- README.md (where each result of the article comes from and how to reproduce it), MANIFEST_SHA256.txt (SHA-256 and size of every file in every zip), EDITED_FILES_SHA256.txt, CITATION.cff, requirements.txt, LICENSE, LICENSE-DATA. The simulation engine and the crowd-voted loop build on the author's earlier release (doi:10.5281/zenodo.20676802, version 2.1.2); extract its code/, data/ and analysis/ folders next to the extracted zips of this deposit. Python 3.14 with numpy, scipy, pandas, matplotlib, mujoco and the Robotarium Python simulator (requirements.txt). The Nav2 episodes need ROS 2 Jazzy and Nav2; the hardware runs need the Robotarium testbed. Code: MIT. Data, results and figures: CC BY 4.0.