arXiv subjects

Dan Popescu

Publications and source records attributed to Dan Popescu.

3 recordsLinked to original sources

arXiv2019

IoT-Enabled Distributed Data Processing for Precision Agriculture

Large scale monitoring systems, enabled by the emergence of networked embedded sensing devices, offer the opportunity of fine grained online spatio-temporal collection, communication and analysis of physical parameters. Various applications have been proposed and validated so far for environmental monitoring, security and industrial control systems. One particular application domain has been shown suitable for the requirements of precision agriculture where such systems can improve yields, increase efficiency and reduce input usage. We present a data analysis and processing approach for distributed monitoring of crops and soil where hierarchical aggregation and modelling primitives contribute to the robustness of the network by alleviating communication bottlenecks and reducing the energy required for redundant data transmissions. The focus is on leveraging the fog computing paradigm to exploit local node computing resources and generate events towards upper decision systems. Key metrics are reported which highlight the improvements achieved. A case study is carried out on real field data for crop and soil monitoring with outlook on operational and implementation constraints.

eess.SP↗

arXiv2019

WSN and Fog Computing Integration for Intelligent Data Processing

Networked embedded systems endowed with sensing, computing, control and communication capabilities allow the development of various application scenarios and represent the building blocks of the Internet of Things (IoT) paradigm. Traditional data collection methods include multiple field level IoT systems that can relay data stemming from a network of distributed ground sensors directly to a cloud platform for storage, analysis and processing. In such applications however, rapid sensor deployment in unstructured environments represents a challenge to the overall robustness of the system. We discuss the fog and mist computing approaches to hierarchically process data along its path from source to destination. The several stages of intermediate data processing reduce the computational and communication effort in a gradual manner. A three-layer topology for smart data monitoring and processing is thus proposed and illustrated to improve the information to noise ratio in a reference scenario.

cs.DC↗

arXiv2018

Erg(r)odicity: Hidden Bias and the Growthrate Gain

Many single-cell observables are highly heterogeneous. A part of this heterogeneity stems from age-related phenomena: the fact that there is a nonuniform distribution of cells with different ages. This has led to a renewed interest in analytic methodologies including use of the "von Foerster equation" for predicting population growth and cell age distributions. Here we discuss how some of the most popular implementations of this machinery assume a strong condition on the ergodicity of the cell cycle duration ensemble. We show that one common definition for the term ergodicity, "a single individual observed over many generations recapitulates the behavior of the entire ensemble" is implied by the other, "the probability of observing any state is conserved across time and over all individuals" in an ensemble with a fixed number of individuals but that this is not true when the ensemble is growing. We further explore the impact of generational correlations between cell cycle durations on the population growth rate. Finally, we explore the "growth rate gain" - the phenomenon that variations in the cell cycle duration lead to an improved population-level growth rate - in this context. We highlight that, fundamentally, this effect is due to asymmetric division.

q-bio.QM↗