A Pilot-Real-Calibrated Indoor Robotic IoT Benchmark Dataset for Edge-Assisted Mobile Robot Navigation and Anomaly Detection

This record provides a pilot-real-calibrated robotic IoT benchmark dataset package for indoor mobile robot navigation, edge-assisted processing, and anomaly detection research. The dataset package was initially designed as a synthetic multi-scenario robotic IoT benchmark and was subsequently calibrated using pilot real-world measurements collected from a TurtleBot3 Burger mobile robot, an LD08 LiDAR sensor, and a Jetson Nano edge device. The package includes calibrated tabular data, metadata files, real pilot evidence files, ROS/ROS2-oriented data collection scripts, Jetson Nano edge-metric collection scripts, validation outputs, and documentation files. The primary data record is: data/calibrated/merged_dataset_pilot_real_calibrated_with_jetson.csv The dataset includes the following major data categories: - robot state and odometry-derived features- LiDAR-derived indoor navigation features- edge computing metrics from Jetson Nano- network and synchronization metrics- scenario labels and anomaly severity values- device, scenario, sensor, and data dictionary metadata- real pilot provenance files and calibration summaries The benchmark covers ten robotic IoT scenarios: 1. normal_navigation2. obstacle_near3. network_delay4. packet_loss5. cpu_stress6. lidar_noise7. multi_robot_interference8. emergency_stop9. low_battery10. wheel_slip Important ethical and methodological note: This dataset should not be interpreted as a fully real-world multi-scenario experimental dataset. It is a pilot-real-calibrated benchmark dataset. The multi-scenario benchmark data were generated synthetically using physically constrained and scenario-specific rules, and selected distributions were calibrated using real pilot measurements from TurtleBot3 Burger, LD08 LiDAR, and Jetson Nano. The real pilot evidence files are included separately in the package to support transparency and reproducibility. Potential uses include: - binary anomaly detection- multi-class robotic scenario classification- anomaly severity prediction- edge-computing performance analysis- ROS-to-tabular feature engineering- benchmarking lightweight machine learning models for robotic IoT systems The dataset is intended to support reproducible research in robotic IoT, indoor mobile robot navigation, edge-assisted robotics, and anomaly detection.