A5000612330- Cryptography and Data Security
- Privacy-Preserving Technologies in Data
- Complexity and Algorithms in Graphs
- Security in Wireless Sensor Networks
- Distributed Sensor Networks and Detection Algorithms
- Stochastic Gradient Optimization Techniques
- Security and Verification in Computing
- Cryptographic Implementations and Security
- Smart Grid Security and Resilience
- Neural dynamics and brain function
- Target Tracking and Data Fusion in Sensor Networks
- Distributed systems and fault tolerance
- Nonlinear Dynamics and Pattern Formation
- Distributed Control Multi-Agent Systems
- Gene Regulatory Network Analysis
- Advanced Data Storage Technologies
- Blockchain Technology Applications and Security
- Chaos-based Image/Signal Encryption
- Genetic and Kidney Cyst Diseases
- Statistical Methods and Inference
- Renal and related cancers
- Sparse and Compressive Sensing Techniques
- Logic, Reasoning, and Knowledge
- Wireless Communication Security Techniques
- Higher Education Governance and Development
Duality (United States)
2023-2025
University of Maryland, College Park
2021-2022
Babeș-Bolyai University
2022
University of Pennsylvania
2016-2020
California University of Pennsylvania
2019
Universitatea Națională de Știință și Tehnologie Politehnica București
2013
Cloud computing and distributed are becoming ubiquitous in many modern control areas such as smart grids, building automation, robot swarms, intelligent transportation systems. Compared to "isolated" systems, the main advantages of cloud-based systems resource pooling outsourcing, rapid scalability, high performance. However, these capabilities do not come without risks. In fact, involved communication processing sensitive data via public networks on third-party platforms promote (among...
This paper explores the privacy of cloud out-sourced Model Predictive Control (MPC) for a linear system with input constraints. A client sends her private states to who performs MPC computation and returns control inputs. In order guarantee that can perform this without obtaining anything about client's data, we employ partially homomorphic cryptosystem. We propose protocols two cloud-MPC architectures: client-server architecture two-server architecture. first case, is privately computed by...
This article develops a cloud-based protocol for constrained quadratic optimization problem involving multiple parties, each holding private data. The is based on the projected gradient ascent Lagrange dual and exploits partially homomorphic encryption secure communication techniques. Using formal cryptographic definitions of indistinguishability, shown to achieve computational privacy. We show implementation results discuss its complexity. conclude this with discussion privacy notions.
Distributed systems are ubiquitous in present-day technologies like smart cities. Such applications require decentralized control, which reduces the load on a single central party, but requires communication and data sharing between participating agents. However, agents might not trust their peers with private data. We propose secure multi-party computation schemes that ensure of control updates each agent, without leaking any other information about states controls neighbors. To this end,...
Real-world healthcare data sharing is instrumental in constructing broader-based and larger clinical datasets that may improve decision-making research outcomes. Stakeholders are frequently reluctant to share their without guaranteed patient privacy, proper protection of datasets, control over the usage data. Fully homomorphic encryption (FHE) a cryptographic capability can address these issues by enabling computation on encrypted intermediate decryptions, so analytics results obtained...
We consider the problem of implementing a Linear Quadratic Gaussian (LQG) controller on distributed system, while maintaining privacy measurements, state estimates, control inputs and system model. The component sub-systems actuator outsource LQG computation to cloud encrypt their signals matrices. encryption scheme used is Labeled Homomorphic Encryption, which supports evaluation degree-2 polynomials encrypted data, by attaching unique label each piece data using fact that outsourced known...
Control as a Service (CaaS) is becoming reality- particularly in the case of building automation and smart grid management. Often, control algorithms CaaS focus on controlling client's system directly from input-output data, since system's model might be private or unavailable. Therefore, large quantities data collected client need to uploaded cloud server. This can used by malevolent service provider infer sensitive information about mount attacks. In this paper, we co-design solution that...
As large amounts of data are circulated both from users to a cloud server and between users, privately aggregating the shared is critical. This article considers problem <i>private weighted sum aggregation with secret weights</i>, where an aggregator wants compute local some agents. Based on privacy requirements posed weights, there different secure multiparty computation schemes exploiting knowledge structure. First, we review for when each agent has private value weight, agents have value,...
This work proposes a protocol for privately solving constrained quadratic optimization problems with sensitive data. The problem encompasses the private data of multiple agents and is outsourced to an untrusted server. We describe desired security goals investigate information leakage from duality theory. present interactive that achieves solution strictly convex linear cost inequality constraints, by making use partially homomorphic cryptosystems securely effectuate computations. Then, we...
The least squares problem with ℓ <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</inf> -regularized regressors, called Lasso, is a widely used approach in optimization problems where sparsity of the regressors desired. As motivation, we investigate sparse data predictive control problem, run at cloud service to system unknown model, using -regularization limit behavior complexity. collected input-output privacy-sensitive, hence, design...
In this paper, we introduce the problem of motion planning with secrecy guarantees. A robot is tracking a desired trajectory, which transmitted on-line by planner, e.g. base station or mobile station. The communication between and planner organized in packets takes place over wireless channel, susceptible to eavesdropping attacks. Our goal design secure codes order encode trajectory information hide it from any eavesdroppers. Meanwhile, should be able recover estimate robot's motion. We...
The development of large-scale distributed control systems has led to the outsourcing costly computations cloud-computing platforms, as well concerns about privacy collected sensitive data. This paper develops a cloud-based protocol for quadratic optimization problem involving multiple parties, each holding information it seeks maintain private. is based on projected gradient ascent Lagrange dual and exploits partially homomorphic encryption secure multi-party computation techniques. Using...
In this paper, we study the problem of jointly retrieving state a dynamical system, as well sensors deployed to estimate it. We assume that possess simple computational unit is capable performing operations, such retaining current and model system in its memory. be observable (given all measurements sensors), ask whether each subcollection can retrieve underlying physical remaining sensors. To end, consider communication between neighboring sensors, whose adjacency captured by graph. then...
We provide an efficient and private solution to the problem of encryption-aware data-driven control. investigate a Control as Service scenario, where client employs specialized outsourced control from service provider. The privacy-sensitive model parameters client's system are either not available or variable. Hence, we require provider perform in privacy-preserving manner on input-output data samples client. To this end, co-design scheme with respect both performance privacy specifications....
Data aggregation in distributed networks is a critical element Internet of Things applications ranging from smart grids and robot swarms to medical monitoring over multiple devices data centers. This paper addresses the problem private weighted sum aggregation, i.e., how ensure that an untrusted aggregator able compute only users' local data, with proprietary weights. We propose scheme achieves confidentiality both weights, as long there are at least two participants do not collude rest. The...
Cloud computing and distributed are becoming ubiquitous in many modern control systems such as smart grids, building automation, robot swarms or intelligent transportation systems. Compared to "isolated" systems, the advantages of cloud-based are, particular, resource pooling outsourcing, rapid scalability, high performance. However, these capabilities do not come without risks. In fact, involved communication processing sensitive data via public networks on third-party platforms promote,...
This paper pertains to the analysis and design of decentralized estimation schemes that make use limited communication. Briefly, these equip sensors with scalar states iteratively merge measurements state other be used for estimation. Contrarily commonly distributed schemes, only information being exchanged are scalars, there is one common time-scale communication estimation, retrieval system achieved in finite-time. We extend previous work a more general setup provide necessary sufficient...