Consistent Geometric Estimation based on a World Model describing Logical Relationships and Sensor Interpretation

Andreas Schierl, Alwin Hoffmann and Wolfgang Reif

When working with multiple independent mobile robots, each has a different knowledge about its environment, based on its available sensors. This paper proposes an approach that allows working with these different views by independently modeling the common logical relationships between the elements in the scene and the meaning of device-specific sensor data. Using these models, for each robot an estimation process can automatically be derived. This process combines and processes the available information to fill in the unknown geometric relationships between the elements and reacts to changes in the logical relationships. The paper presents two different ways of deriving the estimation process, one of which configures a standard unscented Kalman filter. The proposed approach facilitates the consistent coordination of the robots through an application that works on a common world model, while for execution each robot uses its available data and estimations. The approach is demonstrated in two case studies: one with two cooperating mobile robots tracked with an optical tracking system that hand over an object while passing each other, and one with a more challenging sensor constellation.
published 23.12.2017 Journal of Software Engineering for Robotics

For questions regarding the publication, please contact!