Mathematics
2024, 12, 571
2 of 26
(GRU) models, evaluating its performance with the “Edge-IIoTset” dataset [
4
], optimizing
it for resource efficiency, and benchmarking it against existing solutions. We evaluate the
ensemble model’s efficacy in enhancing detection accuracy, its performance using compre-
hensive datasets, its feasibility in resource-constrained environments, and its adaptability
to evolving cyber threats.
The major contributions of this study are as follows.
1.
Innovative Ensemble Architecture: We introduce a cutting-edge model merging CNNs,
LSTMs, and GRUs, harnessing their combined strengths for nuanced intrusion detection.
2.
Use of Real-World Datasets: Our approach is validated using authentic datasets,
ensuring practical applicability in IoT EVCS environments.
3.
Advanced Data Processing Techniques: Sophisticated preprocessing techniques are
employed to manage complex IoT security data, enhancing model learning efficiency.
4.
Comprehensive Performance Analysis: Our model outperforms existing IDS solutions
in accuracy and resilience, proven through extensive testing.
5.
Practical Implications and Scalability: Designed for real-world IoT applications, our
model’s scalability and adaptability offer significant cybersecurity advancements.
6.
Benchmark for Future Research: Setting a new standard in IoT security, our work
paves the way for future innovations in ensemble and hybrid model applications.
In this article, we present a groundbreaking approach to intrusion detection tailored
to the unique challenges of IoT-based EVCS. Through significant advancements in IoT
cybersecurity, we demonstrate the effectiveness and viability of an ensemble model in this
vital domain.
In the remainder of this article, we comprehensively explore network IDS (NIDS) in
the context of IoT-based EVCS. Section
2
presents the nuances of IoT in EVCS, covering the
challenges, cybersecurity threats specific to IoT-based EVCS, and the critical role of NIDS
in safeguarding them. It also highlights recent technological and scientific advancements,
setting the stage for future research directions. Section
3
presents a review of related studies,
providing a scholarly context for our research. In Section
4
, we introduce our proposed
NIDS framework for IoT-based EVCS, detailing its architectural overview, the integration
of CNN, LSTM, and GRU models, data preprocessing techniques, evaluation metrics, and
implementation specifics. Section
5
presents our experimental results, including binary, six-
class, and fifteen-class classification outcomes. Next, in Section
6
, we discuss these results
and interprete their implications. Finally, Section
7
concludes this study, summarizing our
contributions and envisioning the impact of our work in the realm of IoT security.