
(1) Background: Recovered unmanned aerial vehicle (UAV) flight artifacts are increasingly used as evidence, yet recovery is still reported in binary terms, with no quantitative statement of how useful the recovered data are for the question at hand. (2) Methods: We incorporate four intrinsic data-quality dimensions of ISO/IEC 25012, measured following ISO/IEC 25024: Accuracy (here recovery fidelity against the pre-damage log), Temporal Integrity, Completeness, and Consistency, each with an explicit measure in [0, 1] and stated boundary rules. They are aggregated into a Fitness-for-Purpose Index (FFPI) whose weights, aggregation rule, and minimum-dimension floors are set for each evidentiary purpose and which is mapped to three provisional quality bands whose cut-points are not calibrated; the bands are descriptive labels, not admissibility determinations. (3) Results: On a released simulation over the structure of a 105-flight corpus that can no longer be measured, whose platform levels are inputs, the weights affect the assignment of bands only around a cut-point, an accuracy gate removes sets that the weighted mean keeps, and two purpose configurations assign different bands to the same scores. On public data, including 1560 damaged copies of 54 PX4 logs and five DJI flights, damaged in a crossed design and recovered with standard tools, Completeness carries most of the damage (0.978 to 0.789 from 5 to 50 lost global navigation satellite system (GNSS) records), and the other dimensions change little. The simulation overestimates degradation by 0.185 on the composite (95% limits of agreement −0.043 to +0.413), and the cross-device Consistency depends on clock alignment (0.74 to 0.96 after removing a 1.40 s offset) much more than on the matching window. (4) Conclusions: A decomposed, purpose-configured quality statement can be calculated on real recovered logs, and it contains information a recovery percentage does not. Casework use requires calibration against the judgement of the examiner and a new corpus with an independent positional reference.
Deep models report accuracies at or above 99% on the 5G-NIDD benchmark, yet much of that headline reflects how the data are prepared. We show that two forms of leakage (a random split that mixes capture sessions, and identifier-like features that stand in for the label) inflate results, rebuild the protocol so neither can help, and rerun the reproducible baselines under both protocols. On the clean benchmark we introduce LEAF, a lightweight per-flow attention model whose feature gate and attention pool double as inherent explanations and which performs no intra-window mixing because every mixer we tried lowers macro-F1. LEAF is first on accuracy (98.99%), macro-accuracy, and Matthews correlation coefficient (0.985), and statistically second by a small margin on macro-F1 (97.92%). It comes with inherent explanations about 1800× faster than SHapley Additive exPlanations (SHAP), federated training that stays near 99% on non-identically distributed clients and resists Byzantine poisoning to 25% malicious clients with robust aggregation, and lossless 8-bit integer quantization that cuts memory threefold, to about 29 KB, with sub-millisecond inference. We also report where LEAF loses, namely on cross-extractor transfer, on a novel control-plane attack, and under gradient input perturbations.
Security operations require effective prioritization of malware samples that warrant deeper investigation. This study evaluates whether five Portable Executable (PE) metadata features—Packed, Entropy, Packer, Compiler, and Linker—can support lightweight vulnerability-aware triage using PEVuln. After hash-based deduplication, 35,800 unique samples remained, including 559 vulnerability-associated samples. Leakage-safe pipelines evaluated a prior baseline, logistic regression, Random Forest, Extra Trees, and LightGBM under random, family-disjoint, and chronological holdout protocols. Hyperparameters were selected by five-fold cross-validation using Average Precision (AP), thresholds on separate validation data using F2, and uncertainty by 1000 stratified-bootstrap repetitions. Under random evaluation, tree ensembles achieved AP of 0.253–0.255, about 16 times prevalence, while a 5% Random Forest review budget recovered 50% of positives. The highest test AP fell to 0.118 under family-disjoint evaluation and 0.015 chronologically. Feature analyses showed strong complementarity: Entropy alone achieved AP 0.0212, yet removing it reduced full-model AP from 0.2553 to 0.1695. A matched cohort comparison demonstrated superior predictive performance for EMBER, while the five-feature representation retained useful ranking capability at a 5% review budget under both random and family-disjoint evaluation. This representation provides useful first-stage ranking signal but limited robustness under family and chronological shift.
Electric vehicle (EV) charging infrastructure faces security challenges involving unauthorized access, replay attacks, metering-data manipulation, and transaction integrity. This study presents a blockchain-based smart contract framework for charging session management, energy meter authentication, and charging cost computation. The proposed ChargingSessionManager architecture was enhanced with digital signature validation, nonce-based replay prevention, authorized metering updates, and session state transition control. A mathematical security assessment of the system examines authentication, data integrity, replay prevention, and resilience to replay attacks under clearly defined cryptographic and implementation assumptions. It uses a stochastic model to simulate charging load for 200 EVs and 30 charging stations over 60 days, along with behavioral testing of smart contracts in a local Ethereum-compatible environment, providing a strong basis for evaluating the system’s security. The extended evaluation included 102 behavioral tests across three contract deployments, 36 of which were submissions with randomized altered energy inputs. All tests yielded the expected results. The recorded mean gas usage was 177,804.5 gas for initiating the session, 103,065 gas for meter updates, and 79,893 gas for session finalization. These results provide implementation-level proof of the selected security controls in the tested environment, but they do not constitute a comprehensive formal proof, production-level network performance, or defense against all attacks. Compared with those broader benchmarks, the research provides a reproducible basis for further investigation of secure, auditable EV charging management using blockchain-based smart contracts.
Detecting intrusions in database systems is challenging due to the use of legitimate credentials and the scarcity of labeled attack data. This paper presents a specification-based intrusion detection system (IDS) for role-based access control (RBAC) database environments that automatically derives behavioral policies from normal transaction traces and enforces them at runtime. The proposed framework employs a specification check followed by a novel-bigram gate, i.e., a deterministic specification layer that checks role-specific read and write permissions and read-before-write dependencies. An optional novel-bigram gate then flags adjacent operation pairs that never occur in normal training for that role. Isolation Forest was evaluated as a further residual scorer and is not part of the recommended detector. On a synthetic RBAC holdout, the specification layer alone achieves 98.4% accuracy, 100% precision, and 86.0% recall, with zero false positives. The novel-bigram gate raises accuracy to 98.8% and recall to 89.6% while keeping precision at 100%. Isolation Forest detects no further attack and adds eight false positives. These results show that the learned rules and the bigram gate carry the detection and that the tested one-class model does not justify its cost on this corpus. Each detection is accompanied by a human-readable reason, which is useful for audit. The study does not claim efficacy against adaptive adversaries or production DBMS workloads.