Outdoor environment detection processes – Part 2

In part 1 of our brief overview of outdoor environment detection processes, we discussed time-of-flight as well as radar technology. In the following, we now want to take a closer look at ultrasound as well as environment detection with the help of video images. In the final part of our article series, we will discuss laser technology.

Environment Detection with Ultrasound

Detecting the immediate environment surrounding of a vehicle can also be achieved through an acoustic measuring process thanks to ultrasound technology. The frequency of the waves that are emitted, as the name implies, are within the ultrasound range (starting at approx. 16 kHz), which is beyond the audible frequency range of humans. As with radar, distance is calculated based on the duration of the echo. An ultrasound signal is emitted, and then the time required for the sound to return is measured. The model for this method, like many other techniques, comes from the animal world. Bats and dolphins, for example, use ultrasound for navigation. The denser the medium, the faster the sound disseminates. The speed of sound is therefore greater in water than in the air. For this reason as well, ultrasound sensors do not function in a vacuum, since the sound waves cannot propagate there.

In its most simple form, an ultrasound sensor consists of a membrane for sound generation (loudspeaker), a membrane for sound detection (microphone), and a timer. The timer starts once the sound impulse is emitted. Once the reflected sound impulse is detected by the second membrane, the timer stops. Using the known speed of sound, the distance of the reflecting object to the sensor can be calculated.

Today, ultrasound is the predominant sensor for parking assist systems in motor vehicles. The sensor measures the distance to potential obstacles and visualizes them with the help of a display or through acoustic signals. In the meantime, ultrasound sensors are also used in simple lane keeping assist systems in viticulture, or in collision warning systems that detect slow-moving vehicles. The video below shows a test drive with an ultrasound sensor mounted to the side of the vehicle.

Aktivieren Sie JavaScript um das Video zu sehen.

Source: YouTube by MechLabDE

Mono and Stereo Video

Environment detection via video imagery is differentiated into mono video and stereo video. In mono video, a regular video camera like the kind found in webcams or cell phones is used. With the aid of filters and algorithms, objects on the delivered video image are attempted to be recognized. In this way, for example, the gray scale of the camera, which provides especially clear edges, is used to detect a person or a car. The mono camera is thus able to recognize certain objects, but is not (yet) capable of determining the distance to the recognized object. (Structure from Motion may present a possibility in the future.)

In order to also detect distances with the help of a camera, the stereo camera was developed. A stereo camera consists of two cameras that are arranged at a certain interval to one another. Through the recognition of corresponding points in both camera images and the known distance between both lenses, the third dimension, the distance to the objects, can also be deduced through triangulation.

Camera systems are especially popular in the automotive industry since they are relatively economical. Additionally, the existing traffic infrastructure is designed for human visualization. For example, traffic signs or lane markers can only be detected visually, and this still requires the use of sensors that can recognize them.

With performance-capable algorithms and the calculating power they demand, Daimler AG is currently working on a system called 6D-vision that can recognize moving objects such as vehicles or pedestrians, and can also measure their location, directional movement, and speed. This system is briefly introduced in the following video.

Aktivieren Sie JavaScript um das Video zu sehen.

Source: YouTube from Daimler AG

Please also read the other articles of article series concering environment detection:
Part 1
Part 3

Leave a Reply

Your email address will not be published. Required fields are marked *