Assistance sensors for robotics

Roboter der Firma Taurob erkundet brennende Werkstatt.
© Photo Fraunhofer FHR

Robots from the company Taurob investigate a burning workshop.

360° abbildender
© Photo Fraunhofer FHR

360° imaging all-round vision sensor.

Erkundung eines mit Möbeln gefüllten Raumes.
© Photo Fraunhofer FHR

Exploration of a room filled with furniture.

When rescue teams arrive at the scene of an accident, they are often unfamiliar with the site and have no information about the current situation. Radar-based assistance sensors for robots can improve this situation. They provide the rescue teams with a wide variety of information and therefore create a complex image of the accident scene.


What are the application possibilities for radar in the robotics area?

When emergency and rescue teams arrive at the scene of an accident, they have to gather a multitude of information as quickly as possible so as to gain a complete overview of the situation. Each type of incident does, however, require different sensors to assess the situation. A separate robot research area has been set up to address such issues. Fraunhofer FHR contributes radar-based assistance sensors with a broad spectrum of application possibilities.

Due to various combustion processes, a burning laboratory is, for example, quickly filled with smoke with the result that the situation cannot be assessed from a distance. Assessment using robots is an appropriate method of choice. In addition to manual operation on sight, the robotic assistants should, in future, be able to maneuver autonomously and gather the required information independently. Robots that use conventional optical techniques are very limited in their ability to perceive their environment as soot-filled clouds of smoke strongly affect the propagation of the electromagnetic waves in the optical spectrum. The radar-based sensors can generate a variety of information, e.g. a 360° all-round vision or a 3D image of the scene in front of the robot. We are currently addressing both of these themes within the framework of the EU-funded project SmokeBot (Horizon 2020, grant no. 645101).

Why complicated when it can be easy?

When a customer contacts us with a specific problem, Fraunhofer FHR initiates an analysis process to find the simplest solution to this problem. At the same time, we maintain our focus on excellent result quality both with regard to hardware and image quality and always come up with the optimal and most economic technological solution. In the process, we generate high value added for our customers who, thanks to our solutions, can secure a new and stronger position in the market. We adopt this approach in all of our business units and for all of our core competencies, and these include assistance sensors for robots.

The 360° all-round vision of robotic assistants that are capable of navigating in an unknown environment serve as a good example here. In the case of zero visibility referred to above, navigation can once again be made possible using a simple, radar-based sensor solution. This particular solution takes the form of a real aperture-based FMCW radar at 80 GHz, which rotates continuously with 2.5 rotations per second: here, a 360° cross section, which is capable of recording surrounding objects such as furniture and walls even in restricted visibility, is generated around the robot. The enormous bandwidth of the FMCW radar (25 GHz) also facilitates the high-precision measurement of the contours of objects in the cross section. Thanks to this relatively simple concept, the robot is capable of generating information relating to its direct environment autonomously with a high update in practically all environmental conditions (smoke, fog, rain) and independent of daylight. The focus lies on the creation of a simple solution that facilitates the construction of low-cost 360° all-round vision sensors.

3D MIMO with a compact sensor!

Deliberations on a three-dimensional image which, in addition to navigation, would also allow reconnaissance of the objects at the accident scene and even an inspection of the accident scene itself, necessitate a complex solution. Fast image acquisition is of the essence here, with the result that imaging techniques such as real aperture radar (RAR) and synthetic aperture radar (SAR), which are based on mechanical concepts, can no longer be applied. The logical conclusion is the development of a new technology specifically for this field of application.

Fraunhofer FHR is currently conducting research on a FMCW-MIMO array-based sensor at 120 GHz, which allows a fast and three-dimensional assessment of the situation. We are presently in the evaluation phase of the array topology. Due to the wide range of application possibilities, the future clearly belongs to assistance sensors such as these. In addition to implementation in ground-based robots, these can also be integrated into unmanned flying robots and used to gather information on accident scenes which were previous very difficult or even impossible to access.