Cyber-physical systems (CPSs) are highly susceptible to malicious cyberattacks due to their reliance on communication networks. For this reason, many different attack detection techniques have been developed to guarantee the safety of CPSs. This article introduces BlockChain (BC) to address CPS issues such as data security and privacy. Additionally, BC is not well suited for CPS due to its high computing complexity, limited scalability, significant bandwidth overhead, and latency. To meet the requirements of CPS, a light-weight blockchain-based signature algorithm (LWBSA) model is developed in this work. The concept's resource constraints are alleviated by having a single centrally managed manager generate shared keys for outward-bound data transmission requests. The LWBSA paradigm provided herein produces an overlay network where extremely equipped resources can merge into a community BC, hence ensuring both dedicated privileges. Lightweight consensus, the elliptic curve digital signature algorithm (ECDSA), and distributed throughput management (DTM) are the three optimizations implemented in the ELIB model discussed here. Extensive simulation is carried out to examine the implications of different situations on processing time, energy usage, and overhead. The experimental outcomes show that the LWBSA achieves the best possible performance across a wide variety of measures.

The rapid adoption of the Industrial Internet of Things (IIoT) paradigm has left systems vulnerable due to insufficient security measures. False data injection attacks (FDIAs) present a significant security concern in IIoT, as they aim to deceive industrial platforms by manipulating sensor readings. Traditional threat detection methods have proven inadequate in addressing FDIAs, and most existing countermeasures overlook the necessity of validating data, particularly in the context of data clustering services. To address this issue, this study proposes an innovative approach for FDIA detection using an optimized bidirectional gated recurrent unit (BiGRU) model, with the Sailfish Optimization Algorithm (SOA) employed to select optimal weights. The proposed model exploits temporal and spatial correlations in sensor data to identify fabricated information and subsequently cleanse the affected data. Evaluation results demonstrate the effectiveness of the proposed method in detecting FDIAs, outperforming state-ofthe-art techniques in the same task. Furthermore, the data cleaning process showcased the ability to recover damaged or corrupted data, providing an additional advantage.