WP2 Robot Control
Lead Partner: ALU-FR
This WP is concerned with the action the robot can execute on the sense of moving in the environment while
dealing with dynamic scenes, picking and placing actions, conducting inspections, as well as the coordination among robots. These capabilities are essential building blocks for most of the tasks to be fulfilled in this project. In order to robustly move through the environment, a robots needs the following capabilities: It needs means for building a model of the environment, to localize itself in this model, to update it based on changes in the environment, to use the knowledge about the robot’s pose and the state of the environment to plan trajectories towards a desired goal location, and finally to steer free of collisions towards a planned path. These essential navigation capabilities will be developed in the Tasks T2.1 and T2.2 of this WP. It is, however, unrealistic to expect a single robot only. Therefore, a multi-robot framework is developed in Task T2.5 in which a common navigation strategy is defined to avoid collisions and share the sensed data. The exchange of information is essential for optimal behavior and fusing data to build a global map of the environment, which makes the whole process faster, more robust and easily scalable.
In order to pick up objects, a robot needs to recognize and localize objects, plan stable grasps, and collision-free motions. Robustness against sensor noise and computational efficiency are faced challenges in Tasks T2.3. Placing objects in a new position requires accuracy and smoothness. Objects have to precisely placed to their new positions and/or placed smoothly so as not to damage them. In order to achieve this, the robot needs to observe the scene with its sensors and build 3D models of it. The models allow planning and execution of the appropriate object trajectories. This skill will be developed in T2.4.