Site: Institut National de Recherche en
Informatique et en Automatique (INRIA)
Unite de Recherche Sophia-Antipolis
2004, Route des Lucioles - B.P. 93
06902 Sophia-Antipolis Cedex
Date Visited: May 13, 1993
Report Author: M. Lee
Bernard Espiau; Directeur de l'Institut Superieur d'Informatique
Claude Samson; Directeur de Recherche Responsable de Project ICARE
Daniel Simon; Charge de Recherche
Patrick Rives; Charge de Recherche
Institut National de Recherche en Informatique et en Automatique (INRIA) is a coordinating center for scientific cooperation with scientists of the former Soviet Union in the area of underwater vehicle technology. INRIA performs basic research in system design and in control architecture. The institute works closely with IFREMER, which does applied research in underwater vehicles. The concepts that are developed under the ICARE project and that relate to underwater vehicles will be demonstrated on the VORTEX test bed vehicle at IFREMER.
The underwater vehicle and sensor-based control research at INRIA is in the project ICARE (Instrumentation, Control, and Architectures in Robotics), headed by Mr. Claude Samson. This project has fourteen full time people, including six graduate students. The program has three principal investigators: Mr. Bernard Espiau, Mr. Daniel Simon, and Mr. Patrick Rives. Mr. Espiau, who hosted the WTEC team, was scheduled to transfer from INRIA to Grenoble later in 1993.
The focus of this group is basic research in automatic control, and research into programming environments. The objective is the development of a methodology of design, integration, and assembly of many components of a control system (which is able to deal with unknown environments and has real-time programming capability) for complex mechanical systems. These activities are organized into three major projects.
Mr. Simon presented his group's work on sensor-based control. The group is currently using vision as its primary sensor for local control. The scientists are interested in integrating sonar and vision. They would like to use the fusion of this data for local terrain sensing and for a reference for their control system.
The team saw the fixed base robot that INRIA has been using to date for its research platform. A mobile robot platform was demonstrated that will be the institute's future technology test bed. Underwater applications of the technology are done jointly with Mr. Vincent Rigaud at IFREMER.
INRIA is interested in tradeoffs of high speed/low speed control for underwater vehicles. The institute is also interested in characteristics and impact/integration into the control system design for other types of actuators, such as variable ballast systems. The institute's scientists are working on control schemes using time varying control for nonholonomic systems.
Although video servoing is being done at Sophia, the work on video servoing using optical flow is being done at Brest.
One research topic that is being addressed is identifying when the vehicle has enough information to complete a requested task. INRIA is also working on developing an estimation of the errors that evolve when navigating the vehicle from local sensors. INRIA's scientists would like to determine when the sensor drift errors will make completion of the task impossible. This can then be fed into a planning system to define how the vehicle must proceed to get the additional information necessary to complete the task. INRIA is not working on building a large scale world model of its environment. Rather than work from a world model, the institute is focusing on the portion of the world that can be sensed from the vehicle.
The institute's scientists are working on parallel algorithms for planning systems. Their approach is to focus on deterministic algorithms rather than artificial intelligence techniques.
Their first project in control from local sensors started two years ago. The goal of this project was to do bottom following. This application would be used to do a sonar survey and to make maps from the data. Now INRIA is doing work on defining sensors, actuators, and control tradeoffs for the MAST project ARSUAUV (Advanced Research System for Unmanned Autonomous Underwater Vehicles).
The institute is interested in research in underwater intelligent operation. This includes sensor-based control, programming tools, and man/machine interface. The team saw demonstrations of automatic program generation. The inputs to this program are the tasks, the communication links, and the modules that define the servo algorithm to be used. The geometry of the robotic system can be input through a computer assisted design environment. The output is C code that can run in a VXworks environment. INRIA scientists also demonstrated a code simulation system which allows the user to simulate sensor inputs and line-by-line code execution.
INRIA is a basic research organization that is creating technology in software architectures, sensor-based control, and control systems. INRIA is a leader in this area of research. This research is creating enabling technology that will have a significant impact on underwater vehicles. Software architectures that allow easy integration on local sensor-based control will open up the field of underwater experimentation and new methodologies for scientific research using underwater vehicles. It can lower the cost of doing science underwater by reducing the dependency of the vehicles on large scale positioning capabilities.
INRIA has only recently entered the field of underwater robotics and is interested in expanding its role in this area of research.
"Computer-Aided Design of a Robot Controller Handling Reactivity and Real Time Control Issues." December 1992.
Espiau, B., F. Chaumette, and P. Rives. "A New Approach to Visual Servoing in Robotics." IEEE Transactions: Robotics and Automation, Vol. 8, No. 3. June 1992.
Martinet, P., P. Rives, P. Fickinger, and J. Borrelly. "Parallel Architecture for Visual Servoing Applications." Workshop on Computer Architecture for Machine Perception. Paris: 1991.
Rives, P. and R. Pissard-Gibollet. "Reactive Mobile Robots Based on Visual Servoing Approach." AIS 92. Perth: July 1992.
INRIA. Brochure. 1992.