Current projects

• Robotic assistant for persons with disabilities
• Self controlled helicopter
• Dialogue and communication
• Auto 21st
• RobocupRescue
• Robust control for autonomous vehicle fleet
• See past projects

This project, realized jointly by Laval University and McGill University researchers will tackle the problem of designing an intelligent robotic system to help persons with disabilities.

The objectives of this project are to develop an agent able to learn a POMDP model of its own environment and to make both online and offline decisions using this model. The project will also investigate the use spech as the primary mean for the human user to communicate with the agent. Finally, the performance of the developped agent will be evaluated using an existing robot platform (the SmartWheeler) and will be tested with both a control population and users with disabilities.

Participants: Brahim Chaib-draa, Abdeslam Boularias, Hamid Chinaei.

This project works on creating, a drone based on a radio controlled helicopter. The drone shall be able to take off, land, maintain a position and follow some basic trajectories. The helicopter will be modeled as a rational agent, with sensors, effectors, in a partially obserable, stochastic, sequential, dynamic, continuous and multi-agent environment. Techniques such as Kalman filtering and Pegasus algorithm will be used. The ultimate goal of the project is to create a drone able to make its own decisions, use visual recognition and navigate autonomously in an urban environment.

Participants: Brahim Chaib-draa, Maxime Lemonnier.

The intelligent dialogue group studies the co-construction of dialogue in interagents communication. For this research, three kinds of theoretical foundations must be considered:

• theoretical computer science: several communication tools come from information theory, classical computer science or artificial intelligence (protocols, conversation policies, plan-based models of conversation, ...);
• natural language theory: some natural language theories can be applied to artificial systems (Montague's grammar, DRT, LFG, ...);
• agent models: the communication behaviour of an agent is the result of its cognitive model (BDI, BDIO, conflicts avoidance, ...).

With these three dimensions, the intelligent dialogue group explores the interagents communication domain as advanced computer science as well as being a subset of cognitive sciences. Its goal is not the development of systems capable of human-like interactions or of interactions with humans but to consider in interagents communication characteristics of human societies:

• the heterogeneousness of agents: messages must be mutually understandable;
• exchange of knowledge: we must express different kinds of knowledge and not simply data;
• local control: we do not want a central planner nor predefined interactions;
• organizational structures: they are necessary to structure the system and to avoid the combinatorial explosion of the number of transmitted messages.

Participants: Philippe Pasquier , Brahim Chaib-draa, Mathieu Bergeron and Marc-André Labrie. 

The AUTO21 program brings together more than 256 researchers from 42 universities across Canada. It is organized around six research themes, ranging from topics in health and social sciences to materials, manufacturing, fuels and pollution and intelligent systems. There are 41 research projects, each having their leader, as well as three other researchers from at least three institutions in at least two canadian provinces. AUTO21 is sponsored by more than 110 industrial, governmental and institutional partners.

The DAMAS is working on theme F: Intelligent systems and sensors. Our project is called "Vehicle communication and applications" and regroups researchers from Sherbrooke University, Calgary University, Ottawa University and Globis Data Inc..

Participants: Julien Laumonier, Charles Desjardins, Pierre-Luc Grégoire and Brahim Chaib-draa.

This project studies reliability and resilience aspects of Unmanned Air Vehicles (UAVs). Such vehicles, while they become cheaper and more ubiquitous, are still prone to failure. Fault detection and recovery is important to allow the fleet to pursue some high-level mission objectives. Here, at the DAMAS lab, we are interested with robust coordination and planning for a fleet of agents performing a patrol mission amongst a set of locations. This project is a joint project with Concordia University.

Participants: Jean-Samuel Marier and Brahim Chaib-draa.