PhD position (M/F): connected environmental sensors for the study of ecosystems via EURAXESS

Centre national de la recherche scientifique (CNRS)

Castanet-Tolosan, France 🇫🇷

OFFER DESCRIPTION

This research will be carried out on the sites instrumented within the framework of the ECONECT project, and by considering more particularly one of the sentinel systems developed in this project on the aquatic compartment. This system consists in a floating chamber intended to integrate various measuring devices, making it possible to document a certain number of physicochemical parameters (dissolved O2, pH, conductivity, etc.) but also biological metrics such as the chlorophyll content in the water column (providing information on the development of phytoplankton) or the activity of bioindicator organisms.

For the implementation of this project, we will rely on a highly interdisciplinary team, with skills in ecology and ecotoxicology (Arnaud Elger, Laboratory for Functional Ecology and Environment), in design of embedded systems (Jean-Yves Fourniols, LAAS) and in design of communication protocols and architectures (Rahim Kacimi, IRIT).

Scientific and technical context:

The Internet of Things (IoT) opens up particularly interesting perspectives for the study of biodiversity and ecosystems through the use of connected sentinel systems. Such systems are currently being implemented in the ECONECT project, in the Occitanie Region, and will soon be the subject of the structuring project TERRA FORMA, at the national level, the objective of which is to develop the next generation of environmental observatories. While the use of in situ sensors in environmental sciences is not new, there is still a lack of tools to massively collect environmental data to document the coupled dynamics of biotic and abiotic compartments with sufficient spatial and temporal resolution.
The emergence of a new generation of low-cost sensors brings this objective within our reach, but it is also necessary to be able to effectively orchestrate the collection of in situ data and their remote transmission, possibly after a pre-processing of the information as close as possible to the sensors, to transmit only meaningful data.
The solutions developed to be used for example in the context of smart cities or in peri-urban areas are not necessarily usable as such for the study of ecosystems, as the deployment of sensors and network infrastructure in natural environment implies specific constraints. These constraints appear on the energy level, due to the absence of an electricity network and the resulting need for energy sobriety. They also appear in the field of telecommunications, due to the limited or non-existent network coverage. Although many algorithms have been designed for data collection in energy-constrained sensor networks, there are still challenges to be overcome in real deployments (use of the spatio-temporal redundancy of measurements, intermittent links with the base station, congestion induced by traffic convergence in single data-sink networks, etc.).
In addition, the strong heterogeneity of the data collected in environmental issues (measurements of physico-chemical parameters, images and sounds, etc.), particularly in terms of volume, as well as the fact of working at very variable spatial and temporal scales depending on the processes studied, involves the use of different communication protocols due to trade-offs between the range, the data rate and the energy consumption of the different technologies involved. However, being able to offer a versatile and proven architecture, based on a limited number of communication protocols and on simple and inexpensive equipment, would be an important step towards connected and autonomous environmental observatories.

Objectives of the thesis:

We propose to design a versatile communicating electronic architecture based on FPAA and PSOC technologies to converge on reconfigurable analog and digital hardware solutions specifically designed for the collection of environmental data. The technological barrier that we wish to remove concerns the development of an architecture that is sufficiently generic and adaptive to accept different sensors, adapt to environmental variations and detect any degradation of the measurement by minimizing the on-site recalibration steps. Adaptability and the ability to detect “bad measurements” are made necessary by the heterogeneity of the data and the constraints linked to operating in a natural environment in areas that are potentially isolated and / or difficult to access.

We wish to select a limited number of communication protocols and technologies to cover the most common uses for the study of ecosystems and biodiversity. These protocols should allow energy-efficient data transmissions, over relatively short distances, for interactions between sensor nodes at the scale of a study site; but also allow long-range communication to allow data to be routed to the Cloud and remote supervision of on-site devices. In situ characterizations and inter-comparisons of the different communication technologies will make it possible to obtain realistic field data with regard to ranges, data rates and energy consumption in representative environments (i.e. with varied conditions of relief and vegetation cover); these results will make the optimization of the network architecture and the appropriate strategies possible.

To converge on the development of an architecture compatible with a distributed low-power approach and capable of embedding tools for analyzing temporal variations in environmental parameters, the reflection will be conducted in the direction of a harmonization of practices to propose an approach standardized at the hardware and software level, based on proven and low-cost technologies. The processing of the data collected will be optimized, by determining the information to be processed as close as possible to the sensors (via an on-board AI) and that to be processed in the Cloud, with a view to energy and data sobriety. With a view to large-scale deployment, the choice of technologies and components will take into account, beyond the notions of performance and usability, the notions of life cycle, obsolescence, reliability and scalability of the systems provided, in order to guarantee their high level of environmental quality.

Funding of the PhD:

The proposed subject was selected among the Laureates of the 2021 edition of the CNRS 80 Prime scheme. As such, the doctoral student will be funded by a doctoral contract for a period of 3 years. The thesis will be supported by an operating budget (15 k€ in 2021; amount to be defined for the following years). The thesis will take place between the Laboratory of Functional Ecology and Environment (ENSAT site) and LAAS-CNRS.

Desired profile of the doctoral student:

MSc or engineering degree in the field of computer systems, electrical engineering, communication networks or related disciplines, motivated by environmental applications and team work. The candidate must have skills in embedded software development, in particular for the management of a reconfigurable architecture, together with a good proficiency of the C/C++ programming language. Skills in on-board signal processing and AIoT are also desirable. A taste for full-scale experimentation is highly sought after. Good practice of English is required.

Web site for additional job details

https://emploi.cnrs.fr/Offres/Doctorant/UMR5245-ARNELG-003/Default.aspx

Required Research Experiences

  • RESEARCH FIELDEngineering
  • YEARS OF RESEARCH EXPERIENCENone
  • RESEARCH FIELDComputer science
  • YEARS OF RESEARCH EXPERIENCENone
  • RESEARCH FIELDMathematics
  • YEARS OF RESEARCH EXPERIENCENone

Offer Requirements

  • REQUIRED EDUCATION LEVELEngineering: Master Degree or equivalentComputer science: Master Degree or equivalentMathematics: Master Degree or equivalent
  • REQUIRED LANGUAGESFRENCH: Basic

POSITION TYPE

ORGANIZATION TYPE

EXPERIENCE-LEVEL

DEGREE REQUIRED

LANGUAGE REQUIRED