The term radar means Radio Detection and Ranging. Radar is a method of radio measurement technology and works with electromagnetic waves in the gigahertz range. Stationary and moving objects can be detected using radar. A transmitter emits a radio wave into the environment. When the radio wave hits an object, the object reflects the radio wave and can be detected by a sensor.

The following information can be obtained from the reflected radio wave:

• the direction to the object,
• the distance of the object,
• the relative speed between transmitter and object from the frequency shift due to the Doppler effect,
• the succession of measurements describes the path distance and the absolute speed of the object,
• with good radar resolution, the object contours can be detected.

Radar technology is used in aviation, medical applications and autonomous driving. A great advantage of radar technology is that weather and light conditions have no influence on the measurement. In contrast, lidar technology, which is also used in autonomous driving, is dependent on weather and light conditions. The abbreviation Lidar stands for Light Detection and Ranging. Similar to the measurement method used by radar, lidar uses a laser beam to detect the surroundings.

Where are the problem points with radar

Objects are generally not in an otherwise empty environment. For example, in autonomous driving on the road, there are numerous objects such as the walls of houses and traffic signs. These objects also reflect the radar signal. In addition, multiple reflections occur. This is the case when the object to be measured first reflects the radar signal onto other objects and these reflect the radar signal back to the sensor. Therefore, a complex post-processing of the received sensor signal is necessary.

Project work to date

A compact radar system was built and evaluated, which can be used mobile and flexible. In addition to a keyboard and a display, the radar system consists of a transmitter unit, a sensor component and a processing component for the received radar signal. The radar signal is in the frequency range of 77GHz. It is continuously radiated in a frequency modulated manner. The radar system has been upgraded to a dual radar system. This improved the spatial detection, similar to the spatial vision of humans with two eyes. For the classification of objects a Convolutional Neural Networks was implemented. Initial results were published as a paper at IEEE and presented at the International Radar Symposium 2021.

Future project work

The existing radar system will be extended to a multi-radar system. In addition to Convolutional Neural Networks, Recurrent Neural Networks will also be used. In order to save computing power, Compressed Sensing methods are to be used for the signal processing of the sensor signal.

Publications

Steven Duong; Daniel Kahrizi; Sven Mettler; Clemens Klöck:
Moving Target Classification with a Dual Automotive FMCW Radar System Using Convolutional Neural Networks, 21st International Radar Symposium, 2021, Berlin

IEEE Paper

If you are interested in participating in our project or if you would like to do a thesis in this field, please contact the contact persons.