RPL-RIDS: A Multidimensional Dataset Generation Framework for Real-Time Intrusion Detection and Impact Perception in RPL-Based IoT Networks

The RPL-RIDS dataset and its supporting framework adopt a clear hierarchical structure to facilitate users in accessing, using, and understanding the content. The detailed folder structure is as follows: RPL-RIDS/ (Root directory, integrating all related files of the dataset and framework) README.md: Core documentation of the project, including project overview, dataset introduction, and citation instructions. RPLAF_framework_experiment/: Experimental related files of the RPLAF framework, providing the experimental environment and case basis for the dataset. Contiki/: Contiki operating system related files, used to build the experimental environment for RPL protocol simulation. RPLAF/: Core files of the RPLAF framework, supporting the implementation of RPL protocol and related attack simulations. specific_case/: Specific experimental cases (e.g., DIS Flooding attack simulation), including simulation configuration, experimental data, and result files. source_code/: Source code of the entire framework, including feature construction, attack verification, and model analysis scripts, which can be directly run for secondary development or result reproduction. feature_data_construction_pipeline/: Pipeline scripts for feature data construction, realizing the full process from raw data to labeled feature data. 1_extract-readable-feature.ipynb: Script for extracting readable features from raw data. 2_forming-strobing-feature.ipynb: Script for constructing strobing features. 3_forming-base-special-temporal-feature.ipynb: Script for constructing basic special temporal features. 4_forming-log-feature_Energy.ipynb: Script for constructing log features related to energy. 4_forming-log-feature_PDR-E2ED.ipynb: Script for constructing log features related to PDR and E2ED. 5_aligning-fusing-feature.ipynb: Script for aligning and fusing multiple types of features. 6_data_labeling.ipynb: Script for labeling the fused feature data (distinguishing normal and attack data). train_model.py: Script for training intrusion detection models based on labeled feature data. attack_effect_verification/: Scripts for verifying the effect of RPL protocol attacks. attack_effect_verification.ipynb: Jupyter Notebook for verifying and analyzing the impact of different attacks on the RPL network. intrusion_detection_and_impact_perception_model_analysis/: Scripts for intrusion detection model training, testing, and attack impact perception analysis. IDS-models.ipynb: Jupyter Notebook for building, training, and evaluating various intrusion detection models. impact_perception_models.ipynb: Jupyter Notebook for analyzing models related to attack impact perception in RPL networks. dataset_demo/: Demo of the dataset, showing the data form and processing effect of each stage in the feature construction process, helping users quickly understand the dataset. 1_origin/: Original raw data (e.g., PCAP files, log files) of the dataset. 2_readable_feature/: Readable feature data extracted from raw data. 3_strobing_feature/: Strobing feature data constructed from readable features. 4_base_special_temporal_feature/: Basic special temporal feature data. 5_log_feature/: Log feature data (including energy, PDR, E2ED related features). 6_aligning_fusing_feature/: Aligned and fused feature data. 7_labing_feature/: Labeled feature data (with normal/attack labels). 8_normal_attack_data_merging/: Merged dataset of normal and attack labeled data, used for model training and testing. dataset_features_description.pdf: Detailed description document of the dataset features, including feature definitions, data types, and statistical information. ### Usage Tips 1. Users can first refer to README.md to understand the overall project and usage steps. 2. The source_code/ directory provides a complete data processing and model analysis pipeline, which can be run in sequence according to the numbering of the scripts. 3. The dataset_demo/ directory provides sample data of each processing stage, which can be used to test the running effect of the script before using the full dataset. 4. The RPLAF_framework_experiment/ directory provides the experimental environment and cases, which can be used to reproduce the experimental process of the dataset.