Truck drivers tend to have a poor overview of the vehicle's immediate surroundings – their blind spots cover not only the area in front of and behind the vehicle, but also to the sides. Therefore, from July 2024, according to the regulation of the European Commission, trucks and buses must be equipped with safety sensors that can detect motorcyclists, cyclists or other vulnerable road users. This will be ensured by innovative laser sensors and nanoradars, on which experts from FIT BUT worked together.
A child hidden behind the back of a vehicle or a cyclist riding next to a turning truck – even such dangerous situations will help drivers detectnew lidar sensor and nanoradar technologies. As part of two TAČR projects, they were developed by experts from FIT BUT together with the company Valeo, which specializes in research, development and production of assistance systems and systems for autonomous driving.
"We can see some objects on the road from a normal camera, but we don't get information about their distance or speed of movement. This will enable a new nanoradar that can see beyond the vehicles in the line of sight. The LiDAR laser sensor will then offer the driver more detailed information about the geometric shape of the object," Peter Chudý from FIT BUT, who participated in the development of both technologies, explains the basic point of safety sensors.
LiDAR (Light Detection and Ranging) sensors measure the distance of objects using the reflection of a laser beam. They are active sensors that emit energy into the environment and measure the amount that is returned. While the first generation of laser sensors had four layers of the scanning beam, the latest ones have more than a hundred. It can recognize even small objects on the road and warn the driver about them. Even the most famous car manufacturers include them in their assistance systems, including Audi, Mercedes and Honda.
But the ScaLa NFL laser sensor is also unique in its construction: "There are two ways to build a sensor - to focus the beam into a geometrically narrow space, where it can measure distant objects. Or to irradiate the entire space at once, thus monitoring the wider scene," describes developer Radek Maňásek from Valeo.
Example of one of the cases of spatial configuration of a collision scene | Autor: Milan Prustoměrský
The data from the sensor then creates a so-called point cloud – an image where each point represents the distance from the reflected location. Special signal processing software will then take care of its interpretation and assignment of a representative 3D object. It uses machine learning and its development was also undertaken by experts from FIT BUT.
The advantage of the new laser scanner is a quick response to very close objects. "With a sensor based on the scanner principle, where energy is concentrated in a small space, in a smaller space we often encounter the insensitivity of sensors that normally measure up to a meter and a half. They are therefore suitable in situations where the car moves at a higher speed - for example on the highway. However, our latest LiDAR can measure distances from 10 cm," adds Maňásek, who worked on the development of the sensor for 3 years.
LiDAR is thus able to recognize whether there is a child or an obstacle standing in close proximity behind the started car. If the sensor is placed in an autonomous car, it can start an automatic start.
Several sensors are placed on the car to cover the entire area around the vehicle - especially blind spots. They are the ones that pose the greatest risk in trucks or lorries, when the driver sits very high.
The second security element, on which researchers from FIT BUT worked together, is a nanoradar for trucks, which has several advantages over lidar or camera systems – it is not sensitive to the weather, so it does not mind fog, rain or darkness. In addition to higher resistance, it can also immediately measure the speed of the traffic participant.
"It can work even in a dynamic environment and heavy traffic. LiDAR or a camera works more or less like our sight and only sees an obstacle, whereas nanoradar sends out an electromagnetic wave that gets under the car or around it and acquires data even on a shadowed object," explains Michal Mandlík, who is in charge of the development of radar technologies at Valeo.
The nanoradar thus perfectly complements the data collected from other sensors – LiDARs and front cameras. "Radar in the signal chain views road users from above and in a 2D view. It identifies the truck as one rectangle with a protection zone and the cyclist next to it as another. If their routes begin to cross, the driver will immediately receive a signal. LiDAR, on the other hand, provides 3D information and is thus able to identify whether there is a trash can, a person or a small obstacle in front of the car," adds Mandlík.
Thanks to the sophisticated system, the driver can thus obtain information from several different sensors. While lidar technology is great for starting the car, nanoradars will enable perfect monitoring of the sides of the vehicle. "Advanced compilation of information from LiDARs, radars and vision sensors to increase the driver's situational awareness of the vehicle's surroundings was created by experts from FIT BUT. Sensory fusion, the output of which are confirmed trajectories of monitored objects, is one example of the transfer of modern aerospace technologies to automotive," adds Peter Chudý.
The technologies were created separately within the framework of two TAČR projects and it is possible to apply them to the car together or separately. "Of course, the automotive industry deals with costs, and in addition to technical aspects, the price of the resulting car will also decide, which is why a lot of attention is paid to the price of the resulting sensor. In general, it can be said that the price of a laser sensor is several times higher compared to a radar sensor," he points out.
"These are two completely different technologies, but they combine into one user function and provide the driver with an excellent overview of what is happening around the car," he adds. It will then be up to the car manufacturer itself, which combination of sensors the car decides to equip – and this also depends on whether it will be a car with a high degree of automation or a completely autonomous vehicle.
