A Secure Data Aggregation Scheme to Traceback Malicious Charging Piles for Large-Scale V2G Data-Sharing Scenarios

Abstract

In large-scale vehicle-to-grid (V2G) data-sharing scenarios, the secure and accurate aggregation of data from charging piles is crucial to optimise orderly charging services for electric vehicles (EVs) and to support demand response or load forecasting. Existing data aggregation schemes often fail to detect outlier sharing-data sent by charging piles compromised by false data injection (FDI) attacks. To address this, we propose a lightweight secure data aggregation scheme that integrates node-level malicious charging piles traceback and isolation with distributed EC-ElGamal encryption. First, charging piles use the Hellinger-distance of shared charging data between adjacent charging cycles to judge whether the piles are malicious, and through iterative row/column cyclic shift operations, every malicious pile is tracebacked and excluded into a single designated group. Second, distributed key shares create a group public key while each pile retains its own secret key, enabling node-level distributed decryption of the aggregated ciphertext via lightweight EC-ElGamal addition and a single Pollard-lambda lookup. Experiments on 18,061 UrbanEV charging piles demonstrate an 81.7%–100% malicious piles excluding ratio, linear convergence (≈0.07 iterations per added pile), and security analysis proves that the proposed scheme has ECDDH anonymity, collusion resistance and differential-privacy immunity.