Published October 5, 2026 | Version v1

CoreSense Inspection Testbed Simulated Dataset

  • 1. ROR icon Universidad Politécnica de Madrid

Contributors

  • 1. ROR icon Universidad Politécnica de Madrid

Description

This dataset contains 200 ROS 2 rosbag recordings from simulation experiments evaluating three multi-drone coordination architectures for autonomous solar panel inspection: a centralised hybrid planner (BINPAT-based), a decentralised auction-based coordinator, and a decentralised CBBA (Consensus-Based Bundle Algorithm) coordinator. Experiments were conducted using the Aerostack2 framework with a multirotor simulator across four fault-injection scenarios — no failure (S1), single drone failure (S2), solar panel anomaly requiring re-inspection (S3), and simultaneous multi-drone failure (S4) — and four swarm sizes (2–5 drones). Each run directory contains the ROS 2 bag file, computed metrics (makespan, path length, utilisation, fairness, auction convergence time), and the mission specification used. Data are organised as {scenario}/{architecture}/{drone_count}/{replicate}

Files

S0_no_failure.zip

Files (9.1 GB)

Name Size
md5:40c8252e2dd7f30e1b57da1111c09583
968.4 MB Preview Download
md5:b13f63162ae64fcb50d742ad59c72272
1.6 GB Preview Download
md5:a5a12313569a3213bffce14c87c1cd46
2.7 GB Preview Download
md5:fdcefeec55e40040dd727724a644a6bd
1.8 GB Preview Download
md5:80a7e8a95c7e657147f37474393777db
2.0 GB Preview Download

Additional details

Funding

European Commission
CoreSense - CoreSense: A Hybrid Cognitive Architecture for Deep Understanding 101070254

Software

Repository URL
https://github.com/CoreSenseEU/aerostack2
Programming language
C++ , Python