The widespread deployment of deep learning models across various applications has raised significant concerns regarding data privacy. Membership inference attacks (MIAs), a major privacy threat, aim to determine whether a specific sample is used during model training, thereby posing significant risks to sensitive information. Most existing MIA methods rely on the model’s final state output, overlooking the process by which the model memorizes training samples. To better exploit model memorization for MIAs, we propose a novel attack method called machine unlearning-based membership inference attack (MU-MIA). The proposed method introduces machine unlearning to incrementally reduce the model’s memorization of specific samples, generating a forgetting trajectory for each sample. The forgetting trajectory is composed of temporal variations in different metrics of the sample during machine unlearning. To distinguish member from non-member samples, we design a BiLSTM-based binary classifier with attention, which captures discriminative temporal patterns within each forgetting trajectory. Moreover, the machine unlearning phase of our attack is conducted under a zero-shot setting, which eliminates the need for any real data during the unlearning process, thereby improving the practicality and generalizability of the attack. We evaluate the proposed MIA method across different datasets and model architectures, and the comparative experimental results show that our method outperforms existing baseline attack methods.

Fault detection in users’ reactive power compensators (URPCs) remains a critical challenge, particularly for general commercial and industrial consumers lacking technical expertise. Undetected URPC malfunctions not only increase electricity costs for users but also aggravate utility power losses. To address this issue, we propose a novel remote fault detection framework that exploits the joint distribution of active load levels and power factors derived from metering data. A vision transformer with a large margin-aware focal model is then employed to effectively classify the operational states of URPCs, using joint frequency distribution matrices as characteristic representations. Unlike conventional approaches, the proposed method relies exclusively on metering data, thereby simplifying deployment and enhancing accessibility for nonspecialist users. This enables timely operation and maintenance of URPCs, reducing electricity costs and improving overall power system efficiency. The effectiveness of the proposed approach is validated through extensive simulations.

With the rapid development of distributed energy, traditional power consumers have been transformed into prosumers who both produce and consume power, and are able to conduct flexible peer-to-peer transactions. Blockchain is a distributed ledger technology suitable for peer-to-peer transactions. However, the single-chain blockchain cannot simultaneously support the network security check and decentralized power transactions. Therefore, a double-chain blockchain and the corresponding peer-to-peer transaction model are proposed with the reorganization of the transaction chain and power chain. The peer-to-peer independent decision-making transactions are achieved through the transaction chain (public chain) to maximize the prosumer's economic interests. The security check and adjusted transaction scheme are achieved through the power chain (alliance chain) with the power network constraint. Based on the Ethereum blockchain development platform, simulations of peer-to-peer transactions in the IEEE 14-node distribution network are implemented. The simulation results show that the designed double-chain blockchain has higher transaction processing efficiency than the single-chain blockchain. The secure and autonomous peer-to-peer transaction of prosumers can be well supported by the proposed method.

This paper has proposed an electric vehicle (EV) route selection and charging navigation optimization model, aiming to reduce EV users' travel costs and improve the load level of the distribution system concerned. Moreover, with the aid of crowd sensing, a road velocity matrix acquisition and restoration algorithm is proposed. In addition, the waiting time at charging stations is addressed based on the queue theory. The formulated objective of the presented model is to minimize the EV users' travel time, charging cost or the overall cost based on the time of use price mechanism, subject to a variety of technical constraints such as path selections, travel time, battery capacities, and charging or discharging constraints, etc. Case studies are carried out within a real-scale zone in a city where there are four charging stations and the IEEE 33-bus distribution system. The effects of real-time traffic information acquisition and different decision targets on EV users' travel route and effects of charging or discharging of EVs on the load level of the distribution system are also analyzed. The simulation results have demonstrated the feasibility and effectiveness of the proposed approach.

An electric vehicle aggregator (EVA) that manages geographically dispersed electric vehicles offers an opportunity for the demand side to participate in electricity markets. This paper proposes an optimization model to determine the day-ahead inflexible bidding and real-time flexible bidding under market uncertainties. Based on the relationship between market price and bid price, the proposed optimal bidding model of EVA aims to minimize the conditional expectation of electricity purchase cost in two markets considering price volatility. Moreover, the penalty cost of the deviation between the bidding quantities is included to avoid large power variation and arbitrage. The conditional expectation optimization model is formulated as an expectation minimization problem with the conditional value-at-risk constraints. Based on the price data in the PJM market, simulation results verify that our model is a decision-making tool in electricity markets, which can help market players comprehend the variants of bid price, expected cost and probability of successful bidding.

In the smart grid paradigm, growing integration of large-scale intermittent renewable energies has introduced significant uncertainties to the operations of an electric power system. This makes real-time dynamic security assessment (DSA) a necessity to enable enhanced situational-awareness against the risk of blackouts. Conventional DSA methods are mainly based on the time-domain simulation, which are insufficiently fast and knowledge-poor. In recent years, the intelligent system (IS) strategy has been identified as a promising approach to facilitate real-time DSA. While previous works mainly concentrate on the rotor angle stability, this paper focuses on another yet increasingly important dynamic insecurity phenomenon-the short-term voltage instability, which involves fast and complex load dynamics. The problem is modeled as a classification subproblem for transient voltage collapse and a prediction subproblem for unacceptable dynamic voltage deviation. A hierarchical IS is developed to address the two subproblems sequentially. The IS is based on ensemble learning of random-weights neural networks and is implemented in an offline training, a real-time application, and an online updating pattern. The simulation results on the New England 39-bus system verify its superiority in both learning speed and accuracy over some state-of-the-art learning algorithms.

Visible light communications (VLC) with light emitting diodes (LEDs) and photodiodes enables a wide range of applications in LED lighting systems. One of the main challenges in VLC is the limited range of reliable communications, owing mainly to the non-coherent nature of optical communications and limited output power of LED luminaires. In order to enhance VLC range without compromising the network throughput, we consider full-duplex relay VLC. Specifically, we consider a basic 3-luminaire (source, relay and destination) linear system topology under amplify-and-forward and decode-and-forward processing at the relay luminaire. Due to full-duplex operation at the relay luminaire, a self-interference signal component arises and is subsequently canceled out as part of the relay processing. We analyze the system performance in terms of bit error rate (BER) and show that multiple orders of magnitude improvement in BER is achieved as compared to direct link VLC under practical indoor conditions.

We consider illumination sensing in a light emitting diode (LED) lighting system based on frequency division multiplexing (FDM). In the FDM scheme, LEDs render periodical illumination pulses at different frequencies with prescribed duty cycles. The purpose of illumination sensing is to identify the illumination contributions of different LEDs at a target location, within a limited response time. With the use of a photosensor at the target location, the problem of interest is to estimate the amplitude of the sensor signal component due to each LED. In this paper, we propose a successive estimation approach exploiting higher order harmonics of the sensor signal. We show that a larger number of LEDs can be supported using the proposed approach, as compared to our previous approach using only the fundamental frequency component, at the same estimation error.

In this work, we consider illumination sensing in a light emitting diode (LED) based illumination system that normally consists of a large number of LEDs. Illumination sensing is used in order to facilitate the control of such system whose complexity, due to the large number of LEDs, can be quite high. In this paper, key requirements, i.e. accuracy and speed, on illumination sensing are described. Furthermore, we present a filter bank sensor structure based on frequency division multiplexing. The design of the filter response is discussed in the context of supporting maximum number of LEDs, while the key requirements on illumination sensing are satisfied. In particular, it is shown that, through the use of the filter with a triangular filter response, a large number of LEDs can be supported in the presence of frequency offsets in a practical range.

We present Nyquist pulse shaping conditions for two- dimensional (2-D) digital communication systems to be free from inter-symbol interference, for general regular sampling grids. We provide examples of 2-D pulse functions satisfying these Nyquist conditions. In particular, we show that a family of 2-D pulse functions that we construct as separable pulse functions from one-dimensional Nyquist-1 pulse functions obey the 2-D Nyquist criterion and moreover include the one with minimum support area in the frequency domain.

In recent years, LED technology emerged as a prime candidate for the future illumination light source, due to high energy efficiency and long life time. In addition, LEDs offer a superior flexibility in terms of colors and shapes, which leads to a potentially infinite variety of available light patterns. In order to create these patterns via easy user interaction, we need to sense the local light contribution of each LED. This measurement could be enabled through tagging of the light of each LED with unique embedded IDs. To this end, we propose a new modulation and multiple access scheme, named as code- time division multiple access - pulse position modulation (CTDMA- PPM): a form of PPM which is keyed according to a spreading sequence, and in which the duty cycle is subject to pulse width modulation (PWM) according to the required lighting setting. Our scheme considers illumination constraints in addition to the communication requirements and, to our best knowledge, it has not been addressed by other optical modulation methods. Based on the proposed modulation method and multiple access schemes, we develop a system structure, which includes illumination sources, a sensor receiver and a control system. Illumination sources illuminate the environment and transmit information, simultaneously. According to our theoretical analysis, this system structure could support a number of luminaries equal to the size of the CDMA codebook times the dimming range.