Lidar specialist Cepton and the Austrian Industry and Testing Alliance, ALP.Lab, have partnered to build lidar-equipped testing environments for studying complicated road driving situations. The companies hope the project will deliver real-world traffic data from complex urban and rural alpine intersections.
As an alliance of automotive suppliers such as Magna and AVL and a partner to scientific organizations like Virtual Vehicle Research, Joanneum Research and Graz University of Technology, ALP.Lab delivers real-life data on complex traffic situations to test the capabilities of the autonomous driving systems, and helps enable solutions to reduce accidents and improve traffic flows.
From narrow, foggy streets covered in snow to cyclists overtaking a tram in the middle of an intersection, Austria, with its mountainous landscape, sees varied road driving scenarios and faces unique challenges in ensuring vehicle and pedestrian safety. Difficult traffic scenarios could pose obstacles for the seamless deployment of autonomous mobility solutions. Compiling a realistic database of different driving scenarios is, therefore, pivotal to the testing and verification of advanced driver assistance systems (ADAS) and autonomous vehicles (AVs) to ensure that they work well in complex environments.
Through the new partnership, ALP.Lab has deployed Cepton’s Helius Smart Lidar system to capture real-time, 3D traffic data to support the training of artificial intelligence (AI) algorithms used for the scenario-based testing of ADAS and AV features. Additionally, the collected data should make it possible to predict vehicle trajectories at intersections and enable traffic data analysis for traffic flow optimization.
As part of the Helius system, a number of Vista-P60 lidar sensors from Cepton have been installed at the corners of road intersections and traffic circles, complementing other sensor types including radar and other types of optical sensors. Cepton claims that for traffic monitoring and analysis applications, lidars perform well in many critical cases where other sensor types struggle. It also notes that lidar offers 24/7 availability thanks to its perception capabilities in various lighting conditions, such as night-time, shadows and the glare of strong headlamps. Lidars also offer higher angular resolution compared with radars and minimize false negatives and false positives.
The P60 lidars, powered by Cepton’s Micro Motion Technology (MMT), are claimed to offer a high detection range of 200m at 30% reflectivity and an angular resolution of 0.25°, making it possible to accurately detect each arriving and departing vehicle and pedestrian at a distance of up to 170m.
The Helius perception software adds an additional layer of intelligence to the lidar data. It takes the point cloud data from a connected network of sensors and delivers real-time, 3D information on object dimensions, locations and velocities. This enables the software to accurately track and classify the different objects involved in a traffic situation. As Helius and the other sensors used by ALP.Lab only capture anonymized data, it also protects the privacy of pedestrians and vehicles while feeding rich traffic data for deeper analysis.
“We primarily selected Cepton´s lidar solutions because of their durability and precise detection capabilities even at high vehicle speeds. With the information the lidar sensors provide, our data is as close to real life as it gets. Therefore, we are able to create enhanced simulation and AV testing environments as well as detailed analyses of the traffic flow,” said Christian Schwarzl, director of testing and validation from ALP.Lab.
“There is amazing synergy between Cepton and ALP.Lab, as we share the same goal of making ADAS and AV features available to the mass market and to improve the safety of vulnerable road users.”