CMOS Sensor Allows Robots to ID Objects in 3-D

A new CMOS sensor implemented in unmanned aerial vehicles (UAVs) gives the aircraft autonomy, improving collision detection as well as enabling 3-D modeling of disaster areas and surveillance targets.

Researchers at the Fraunhofer Institute for Microelectronic Circuits and Systems (IMS) have developed a CMOS distance sensor for UAVs that can efficiently measure 3-D distances. Each pixel in the sensor is given a gray value and a distance value, enabling the UAV to sense where it is in relation to objects around it and to avoid collisions.

The 3-D camera in the flying robot can identify small objects measuring 20 x 15 cm from 7 m away. (Image: Fraunhofer IMS)

According to Werner Brockherde, head of the IMS development department, the sensor has a much higher resolution than radar and can resolve objects of 20 x 15 cm at a range of 7.5 m.

The sensor will also give accurate readings despite the interference of other bright-light sources. The shutter on the camera opens only for a few nanoseconds and takes two measurements. The first is taken using only ambient light; the other uses a light pulse that reflects off of surrounding objects. The sensor analyzes the two inputs and determines the required output in real time.

The sensor has myriad uses, from 3-D modeling of a disaster area for relief workers, to modeling of streets and buildings for urban planners.

Cameras using the sensors have been manufactured by TriDiCam, a spinoff company of Fraunhofer IMS, as part of the AVIGLE project, winner of the Hightech.NRW competition, which is funded by Land of North Rhine-Westphalia and the EU.

The IMS researchers will present their developments at the Fraunhofer CMOS Imaging Workshop in Duisburg on June 12 and 13.

For more information, visit: www.fraunhofer.de/en