Sensor data collected by Internet of Things (IoT) devices can reveal sensitive personal information about individuals, raising significant privacy concerns when shared with semi-trusted service providers, as they may extract this information using machine learning models. Data obfuscation empowered by generative models is a promising approach to generate synthetic data such that useful information contained in the original data is preserved while sensitive information is obscured. This newly generated data will then be shared with service providers instead of the original sensor data. In this work, we propose PrivDiffuser, a novel data obfuscation technique based on a denoising diffusion model that achieves a superior trade-off between data utility and privacy by incorporating effective guidance techniques. Specifically, we extract latent representations that contain information about public and private attributes from sensor data to guide the diffusion model, and impose mutual information-based regularization when learning the latent representations to alleviate the entanglement of public and private attributes, thereby increasing the effectiveness of guidance. Evaluation on three real-world datasets containing different sensing modalities reveals that PrivDiffuser yields a better privacy-utility trade-off than the state-of-the-art in data obfuscation, decreasing the utility loss by up to 1.81% and the privacy loss by up to 3.42%. Moreover, compared with existing obfuscation approaches, PrivDiffuser offers the unique benefit of allowing users with diverse privacy needs to protect their privacy without having to retrain the generative model.

Optimizing the configuration and placement of sensors is crucial for reliable fall detection, indoor localization, and activity recognition in assisted living spaces. We propose a novel, sample-efficient approach to find a high-quality sensor placement in an arbitrary indoor space based on grey-box Bayesian optimization and simulation-based evaluation. Our key technical contribution lies in capturing domain-specific knowledge about the spatial distribution of activities and incorporating it into the iterative selection of query points in Bayesian optimization. Considering two simulated indoor environments and a real-world dataset containing human activities and sensor triggers, we show that our proposed method performs better compared to state-of-the-art black-box optimization techniques in identifying high-quality sensor placements, leading to an accurate activity recognition model in terms of F1-score, while also requiring a significantly lower (51.3% on average) number of expensive function queries.

This article proposes a sensor data anonymization model that is trained on decentralized data and strikes a desirable trade-off between data utility and privacy, even in heterogeneous settings where the sensor data have different underlying distributions. Our anonymization model, dubbed Blinder, is based on a variational autoencoder and one or multiple discriminator networks trained in an adversarial fashion. We use the model-agnostic meta-learning framework to adapt the anonymization model trained via federated learning to each user’s data distribution. We evaluate Blinder under different settings and show that it provides end-to-end privacy protection on two IMU datasets at the cost of increasing privacy loss by up to 4.00% and decreasing data utility by up to 4.24%, compared to the state-of-the-art anonymization model trained on centralized data. We also showcase Blinder’s ability to anonymize the radio frequency sensing modality. Our experiments confirm that Blinder can obscure multiple private attributes at once, and has sufficiently low power consumption and computational overhead for it to be deployed on edge devices and smartphones to perform real-time anonymization of sensor data.

With the rollout of smart meters, a vast amount of energy time-series became available from homes, enabling applications such as non-intrusive load monitoring (NILM). The inconspicuous collection of this data, however, poses a risk to the privacy of customers. Federated Learning (FL) eliminates the problem of sharing raw data with a cloud service provider by allowing machine learning models to be trained in a collaborative fashion on decentralized data. Although several NILM techniques that rely on FL to train a deep neural network for identifying the energy consumption of individual appliances have been proposed in recent years, the robustness of these techniques to malicious users and their ability to fully protect the user privacy remain unexplored. In this paper, we present a robust and privacy-preserving FL-based framework to train a bidirectional transformer architecture for NILM. This framework takes advantage of a meta-learning algorithm to handle the data heterogeneity prevalent in real-world settings. The efficacy of the proposed framework is corroborated through comparative experiments using two real-world NILM datasets. The results show that this framework can attain an accuracy that is on par with a centrally-trained energy disaggregation model, while preserving user privacy.

With the growing adoption of electric vehicles (EVs) and advent of bidirectional chargers, EV aggregators, such as charging stations, will become a major player in electricity markets, providing voltage regulation (VR) or other services. We present a novel and practical VR scheme that takes advantage of the charging flexibility of EVs in charging stations that are connected to buses in a distribution grid. This VR scheme relies on real-time measurements, as well as estimates of the distribution system state and regulation capacity of each charging station. We then propose a novel false data injection attack against the VR capacity estimation process that exploits the uncertainty in EV mobility and network conditions. We show the attack vector with the largest expected adverse impact is the solution of a stochastic optimization problem, subject to a constraint that ensures it bypasses bad data detection. We determine this attack vector by solving a sequence of convex quadratically constrained linear programs. The case studies examined in a cosimulation platform, based on two standard test feeders, reveal the vulnerability of the VR capacity estimation process.

The abundance of data collected by sensors in Internet of Things (IoT) devices, and the success of deep neural networks in uncovering hidden patterns in time series data have led to mounting privacy concerns. This is because private and sensitive information can be potentially learned from sensor data by applications that have access to this data. In this paper, we aim to examine the tradeoff between utility and privacy loss by learning low-dimensional representations that are useful for data obfuscation. We propose deterministic and probabilistic transformations in the latent space of a variational autoencoder to synthesize time series data such that intrusive inferences are prevented while desired inferences can still be made with sufficient accuracy. In the deterministic case, we use a linear transformation to move the representation of input data in the latent space such that the reconstructed data is likely to have the same public attribute but a different private attribute than the original input data. In the probabilistic case, we apply the linear transformation to the latent representation of input data with some probability. We compare our technique with autoencoder-based anonymization techniques and additionally show that it can anonymize data in real time on resource-constrained edge devices.

This paper presents a novel big data analytics framework for creating explainable personas for retail and business banking customers. These personas are essential to better tailor financial products and improve customer retention. This framework is comprised of several components including anomaly detection, binning and aggregation of contextual data, clustering of transaction time series, and mining association rules that map contextual data to cluster identifiers. Leveraging rich transaction and contextual data available from nearly 60,000 retail and 90,000 business customers of a financial institution, we empirically evaluate this framework and describe how the identified association rules can be used to explain and refine existing customer classes, and identify new customer classes and various data quality issues. We also analyze the performance of the proposed framework and show that it can easily scale to millions of banking customers.

The growing adoption of Distributed Energy Re-sources (DER) in low-voltage distribution grids calls for new feedback control algorithms that rely on quasi-real-time data collected by remote sensors. The design and evaluation of such algorithms necessitates a prudent and comprehensive approach since these algorithms require a tight integration of power and communication systems. A simple link failure or a sophisticated cyberattack launched against the grid's monitoring, communication, and control infrastructure could rapidly grow out of control, making the grid unstable. We investigate the design and implementation of a high-fidelity smart grid simulation platform which integrates a network simulator and a power flow simulator using the Mosaik co-simulation framework. The platform allows for evaluating the performance of new control algorithms and understanding dynamics of modern distribution grids. Example case studies are presented to validate the proposed platform.

Topology control algorithms in mobile ad-hoc networks aim to reduce the power consumption while keeping the topology connected. These algorithms can preserve network resources and increase network capacity. However, few efforts have focused on the issue of topology control in presence of node mobility. One of the notable mobility aware topology control protocols is the ldquomobility aware distributed topology control protocolrdquo. The main drawback of this protocol is on its mobility prediction method. This prediction method assumes linear movements and is unable to cope with sudden changes in the mobile node movements. In this paper, we propose a pattern matching based mobility prediction method in which every mobile node predicts its future location through finding similar patterns in its history of movements. Simulation results show significant improvements in terms of prediction accuracy and power consumption compared to the other known algorithms.