Each exercise will consist of a laboratory session. Usually, the lab part will focus on experimental work using simulators or real hardware platforms. The student will have to collect data and sometimes write a few lines of code. The balance between practice and theory will of course be competely dependent on the topic of the lab.
Trail laying and following mechanisms, emphasizing SI concepts; Ant Colony Optimization
Introduction to Webots, a realistic, embodied, sensor-based robotic simulator.
Introduction to the e-puck robot. Illustrate key concepts of the course for basic behavior using different reactive control architectures (Artificial Neural Network, linear Braitenberg, behavior-based, rule-based). Simple localization algorithms based on odometry.
Multi-robot localization, coordinated and collective movements in microscopic model (matlab/point-simulator visualized with Webots)/Webots, includes some collective movement analysis.
Multi-robot systems coordination using market-based and threshold-based algorithms using Webots/Matlab/point-simulator.
Introduction to Mica-z sensor nodes. Simulated (Webots with Omnet++ plugin) and real (e-puck and Mica-z) sensor and actuator networks: networking static sensor nodes with mobile robots for performing collective decisions.
Practical lab verification test, subject: lab 1 to 6.
Multi-level modeling of distributed robotic systems.
Particle Swarm Optimization. Application to benchmark functions and control shaping for single robot (in simulation).
Particle Swarm Optimization application to multi-robot systems (Webots), task obstacle avoidance, tight collaborative task (e.g., formation 2 robots).
Distributed sensing with static, mobile, and robotic nodes (implementation in Webots).
Practical lab verification test, subject: lab 7 to 10.