Task Representation in Robots for Robust Coupling of Perception to Action in Dynamic Scenes

Abstract

Most current perception systems are designed to represent static geometry of the environment and to monitor the execution of their tasks in 3D Cartesian representations. While this representation allows a human-readable definition of tasks in robotic systems and provides direct references to the static environment representation, it does not correspond to the native data format of many of the passive sensor systems. Additional calibration parameters are necessary to transform the sensor data into the Cartesian space. They decrease the robustness of the perception system making them sensitive to changes and errors. An example of an alternative coupling strategy for perception modules is the shift from look-then-move to visual servoing in grasping, where 3D task planning is replaced by a task definition defined directly in the image space. The errors and goals are represented here directly in sensor space. In addition, the spatial ordering of the information based on Cartesian data may lead to wrong prioritization of dynamic objects, e.g., not always the nearest objects are the prime collision candidates in the scene. We propose alternative ways how to represent task goals in robotic systems that are closer to the native sensor space and, therefore, that are more robust to errors. We present our initial ideas, how this task representations can be applied in manipulation and automotive domains.