Technologies in Robotic Systems

Mobile System Position Sensing and Navigation

Position Sensing and Navigation: Key Technology of Service and Assistive Robotics

Reliable localization of mobile systems, route planning and position control are essential to many applications in logistics, manufacturing and service robotics. Examples include item tracking in warehouses, the delivery of resources to assembly lines or an assistive robot’s autonomous navigation and execution of localized actions.

Our know-how in the field of localization and navigation encompasses not only established methods such as positioning with external sensors, laser scanners and artificial landmarks but also novel approaches that operate with inexpensive sensor systems and without modifying the environment.

We develop customized solutions, modifying cost and accuracy for your requirements. Our extensive software framework facilitates rapid integration in an overall scenario and flexible combinations of different systems.

© Fraunhofer IFF
© Fraunhofer IFF
© Fraunhofer IFF

Position Control

The challenge with the autonomous welding robot NOMAD was to position an industrial robot on a mobile platform precisely. An external camera system detects the platform’s position. Then a path is computed and the omnidirectional platform approaches the workpiece being welded.

Laser-based Localization

Autonomous platforms’ self-localization not only serves to determine position but also to avoid obstacles. The use of laser scanners that reliably detect local obstacles has therefore become established in enclosed spaces. Related systems that scan environments two-dimensionally have been researched extensively and function reliably. The drawback is a laser scanner’s inability to perceive more than one cutting plane in an environment. Objects outside the cutting plane are however also relevant in dynamic environments. Thus, three-dimensional systems are necessary. The assistive robot LiSA therefore employs additional laser scanners that are aimed obliquely upward. Thus, obstacles like open drawers and overhanging tabletops are detectable. Other approaches employ time-of-flight cameras or moving laser scanners to scan environments three-dimensionally.

Camera-based Systems

Another drawback when using laser scanners is the high cost of the sensors. A purely camera-based system that localizes a mobile platform was therefore developed in the project AVILUSplus. This visual localization system relies on inexpensive sensors and operates without artificial landmarks. Natural features on the ceiling or walls are mapped and used for self-localization. Additional artificial landmarks can be integrated in the localization process at any time to increase its precision. This system attains accuracies within a few centimeters.

Camera-based navigation and localization of a mobile robot through the use of existing natural features in factory buildings.

Customized Systems

Special applications require customized systems. A camera-based positioning system was implemented when we were developing inspection robots for sewer lines. It determines the inspection systems’ location along the axis of the sewer line with an accuracy of a few millimeters. The principal challenge in the unusual environment of a concrete pipe is the detection of features when one is mainly dependent on fine structures and pores in the surface of the concrete. This system precisely documents the location of damaged spots and finds them again.

Safety

Camera-based systems have been difficult to certify as safety sensors. We are therefore working on inexpensive solutions for households, which simultaneously fulfill a safety function. One option is tactile sensors that can be applied to a robot as an artificial skin. They enable a robot to feel along obstacles or navigate through a cluttered room.