In most robotic applications, robot actions (such as pick, place, push, touch, mount, etc.) are programs created by expert roboticists. In the face of the uncertainties and flexibility need of the real world, these approaches often lack the required robustness. When we consider how humans solve dexterous manipulation tasks, we observe that many modalities (vision, proprioception, touch, etc.) are combined to form a robust reactive procedure for manipulating and navigating our surroundings. In this project, we will design a mapping algorithm that will enable us to create such a multimodal representation of actions that incorporates prior knowledge about the uncertainty of different domains, and show how this can be used to ease teaching robotic actions and generalize them to new embodiments (e.g. different robots, grippers, etc.) and environment. Finally, we will apply the results to learning by demonstration.

- Karla Stepanova    

One of the key goals of collaborative robotics is sharing a workspace between robot and a human operator. The shared workspace is used by both human and robot to complete a common task – either in parallel or as a part of active collaboration. In both cases, the robot should not require separation from the human operator by a protective barrier (e.g., a metal cage). An important aspect of a collaborative robot is the understanding of the scene (i.e. where is the human and what and where are the objects on the scene/workspace) and the ability to react to the immediate needs of the operator (e.g., need for a specific tool). Additionally, the robot should be able to predict the future steps in the production process.

- Radoslav Skoviera