Project description


Project organisation

Project is organized around the following main fourth axis:

  • WP1. Definition of models (for multi-agent systems and communication network structure), and a study case
  • WP2. Multi-agent coordination and control design
  • WP3. Development of a simulator and a graphic interface
  • WP4. Validation via hardware-in-the-loop.

WP1. Definition of models for multi-agent systems and communication
network structure, and specification of a study case.

Definition of adequate models is a key point for the successful design of control strategies. Besides the definition of suitable models for the different types of vehicles, including underwater and surface vehicle. Attention will be paid to the design of models for marine communication including fading and reflection effects. A study scenario for a “gradient search and following for [x] source detection” mission (where [x] is representing the nature of the source to be detected: fresh water, chemical source, methane vent,….) will be defined. The objective of this mission is to detect and follow gradients of concentration of the source flow by considering sensed data provided by dedicated scientific sensors located on-board the AUVs.

WP2.- Multi-agent coordination and control design

Investigation of algorithms enabling the development of a number of Multi-Vehicle Primitives (MVP). Those aimed at achieving Path Following and Trajectory Tracking, Coordinated Path Following, and Cooperative Manoeuvring in the presence of tight communication constraints. Study how to accommodate large communication delays, intermittent communication failures, and temporary loss of navigational data. Particularly relevant will also be the study of algorithms to determine when the vehicles should communicate to interchange knowledge about their state in order to minimise energy expenditure and to reduce the data rates involved. The tools used will borrow dynamical systems theory (control and filtering) and also from the recent field of Networked Control Systems.

The research topics will be concentrated along the following 3 main lines:

1. Coordination and control under computation resource constraints. This includes: formation of fleets, source search, localization algorithms, and management of computing resources,

2. Information & energy management for coordination and control. Here we study: source coding, communication networking, formation of fleets with maximum communication visibility, energy-aware algorithms,

3. Multi-carrier modulation strategies for acoustic marine communication. Adaptation of the multi-carrier modulation strategies for the marine medium with the aim to optimize the bandwidth use and to increase robustness of the transmission.

WP3. Development of a simulator with an interactive graphical interface

The main goal of this simulator with an advanced graphical interactive interface is to validate the fundamental principles developed in previous work-packages. It integrates agent's models, communication media and computation resources including their limitations, heterogeneous network, and all the variants of the multi-agent control strategies.

This simulator is a shared platform where all our findings will be integrated and validated. The validation scenario (marine habitat mapping) is a real-size application complex enough to enforce the pertinence of our results. It will be further used as an open research tool for various applications in the field of multi-agents networked systems.

Besides the models and simulation engine the simulator is complemented with an interactive graphical interface which is used to visualize and interpret the state of the multi-agent control system in one direction and to send high or low-level controls to the agents.

WP4. Simulation validation via Hardware-in-the-loop

The AUV simulation platform consists in a complete set-up reproducing the hardware and software environment of the real AsterX AUV owned and operated by IFREMER. The platform is built around three main computers having different roles:

  • the mission programming computer allowing the definition of the mission to be achieved by the vehicle,
  • the surface controller connected to the vehicle prior to the vehicle launch (vehicle on the deck) or when the vehicle is at the surface (communication using radio communication link) and,
  • the vehicle controller in charge of executing the downloaded mission in an autonomous way (during the mission execution, no more communication with the surface controller is possible).

A second simulation platform developed by Ifremer allows to simulate and model acoustic equipment used either for positioning and transmission purposes. Also, physical environment characteristics can be modelled in order to produce realistic conditions to test the acoustic equipment.

The connection of the two simulators (the simulator produced within the project and the Ifremer AUV simulator) will allow to envisage a global simulation where the AUV is executing its mission meanwhile having the possibility of performing acoustic communication with an external system : surface vehicle, ship or another sub-sea vehicle. In this scenario, the Ifremer AUV simulator will play the role of one of the elements (node) of the coordinated systems network of the project simulation. This will validate the various approaches and model developed and used in the project simulation for managing dynamic controlled networks.

As an option for the project, Ifremer will propose to replace the AUV simulator with the real AUV operated under experimental conditions at Ifremer facilities.