2020 is the year when the first Estonian autonomous shuttles by Auve Tech are introduced on both closed areas and public roads, making transit and discovery of large areas more accessible and more enjoyable. Tallinn Zoo and Estonian Open-Air Museum introduced its visitors to Auve Tech’s autonomous shuttle bus that makes discovering large areas more enjoyable. The vehicle can move around safely because of EyeVi’s unique mapping system that sped up time to market and is based on a reasonable, scalable business model.
Traditionally maps have been made by flying over vast areas of land by plane and taking aerial photos while centimetre accuracy features are collected by geodesist on the spot. EyeVi combines these two and it can extract engineering-grade feature-rich datasets with the help of its unique software platform while driving through vast areas.
With its high-definition mapping technology, the Estonian-based start-up has brought cartography to a whole new level, that is also where autonomous vehicles can most benefit from it. „We are happy about the cooperation with Auve Tech in helping them to kick-start the next step of innovation in Tallinn Zoo, Estonian Open-Air Museum, Ülemiste City and Tallinn Airport,” says EyeVi’s founder and CEO Gaspar Anton.
He explains that to EyeVi, this cooperation entails much more than just software as a service delivery. „We are evolving with our client and understanding the issues autonomous shuttle developers are facing in terms of the usage of 3D data and high-definition mapping,” Anton recounts. He underlines that this is where he sees that their company’s decade-long-experience is valued.
Combines different features
„For autonomous vehicles to navigate, many of them need some sort of baseline data on the current situation,” explains EyeVi’s lead engineer Kalev Julge, while describing how their solution works. There are different ways to collect this data. EyeVi’s mobile mapping system combines 360 imagery and 3D LIDAR data with AI-based feature extraction processes.
„Our solution aims at providing a point cloud of an area where the vehicle is supposed to drive. With its integral LIDAR device, it measures these points and navigates within their limits,” Julge explains. By using point cloud vector maps the vehicle sees where different objects such as driving lanes, stop signs, street lighting, etc. are.
During the data collection process, EyeVi’s mobile mapping system is mounted on a car. The technology combines a panoramic camera, a LIDAR scanner and GNSS/INS devices that enable capturing high accuracy geodata. Collected data is combined and with the help of EyeVi developed unique software a point cloud is formed.
A point cloud is a combination of points, where each point has an X-, Y- and Z-coordinate as well as the intensity. „The intensity depends on how intensely the beam reflects from a point,” tells Julge. Intensity enables the autonomous vehicle to distinguish road surface markings from adjacent objects because the road markings are white and reflect more strongly.
Has offered some great advice
The high definition mapping solution that EyeVi provides includes data collection, feature extraction and an end user application for data management. This is also how the high accuracy point maps used for navigating Auve Tech’s autonomous shuttles in the closed areas of Tallinn Zoo and Estonian Open Air Museum were born in the beginning this year and have now resulted in the piloting of the shuttles on public roads in Ülemiste city business district.
Even though, in many cases, autonomous vehicles already have all the necessary equipment to produce a high-definition map, what they usually lack is the software platform that EyeVi can provide them with. This was also the case with Auve Tech.
„As autonomous vehicles need 3D maps of the areas where they are going to be operating in, we were searching for a company that could do that for us. EyeVi is the only one who can do it in Estonia,” tells Auve Tech’s chief software engineer Ott Männik. According to him, EyeVi’s team has been very flexible.
„They always meet our wishes, and when providing us with high-definition maps, they use the same format as we do,” he adds. Auve Tech got some great recommendations from EyeVi, for example, on how to connect the LIDAR they have on their autonomous shuttle to a GPS-system. „They’ve given us some great sensor-related advice,” affirms Auve Tech’s chief software engineer.
The company itself is also an Estonian start-up. It aims to change the future of mobility by developing and manufacturing autonomous electric and hydrogen vehicles that use LIDARs, sensors, and cameras to react to its surroundings. Their autonomous shuttle bus that was available for the visitors of Tallinn Zoo this spring made it easier to commute in the vast areas of the zoo.
Beside Auve Tech’s autonomous vehicles, EyeVi has gathered point clouds for other companies both in Estonia and abroad. In summer 2019, EyeVi served as a subcontractor for American-based Carmera that provides real-time high-definition maps and navigation-critical data for the world’s leading developers of autonomous vehicles.
Novel business model
What makes EyeVi distinct from other similar players in the market is its scalable business model and flexibility in the execution. The mobile mapping technology it uses can be attached basically to any vehicle; there is no need for a specially designed car to do the job. As EyeVi’s technological set-up is very cost-effective, large areas can also be covered with more than one vehicle.
EyeVi’s business model centres around the idea that most of the autonomous vehicles are already equipped with necessary sensors to produce high-definition maps. The software EyeVi provides is sensor agnostic. It can adapt to the vehicle and integrate its specific dataset while building a 3D point cloud and base data for a high-definition map.
Scalability comes from the fact that the royalty fees are paid according to the project, i.e., in case of a smaller project, the fees are low as well. Accompanied by its ten-year-experience EyeVi can help any autonomous vehicle developer understand the benefits and risks of using map-related data specifically on their vehicle and guide them in the right direction.