Centimeter-accurate localization and mapping in real time
Visual Simultaneous Localization and Mapping (vSLAM) allows for intelligent environment perception in volume products on a cost-effective and scalable basis. It can also be used to support or even replace some LiDAR systems in highly automated vehicles. To this end, IAV has developed a dedicated algorithm, focusing on applications in the mobility sector.
The so-called vSLAM technology has already proven its worth in numerous applications. For example, it is used for the navigation of autonomous drones or for suction robots to help them find their way around the home. Unlike lidar-based systems, which use laser beams to measure the distance to walls and objects, vSLAM relies on information from a camera installed on the robot – similar to the spatial perception of the environment by the human eye. With the help of this information, software constructs an always up-to-date three-dimensional map of the room, within which a robot can determine its exact position and react to changing conditions.
vSLAM offers many more promising applications for intelligent environmental perception: High-precision localization is a cornerstone of intelligent environmental perception. Image recognition technologies and AI algorithms can often only show their full strength if they know where an object is in relation to a vehicle or robot. With vSLAM, applications such as trained parking or AR navigation can be made available to a broad mass.
«Highly precise localization is a cornerstone of intelligent environment perception.»
— Expert in image processing and localization
As most volume-produced vehicles are already equipped with the necessary camera hardware, vSLAM is a purely software-based solution. It is capable of providing a cost-efficient and yet extremely effective addition to Radar and ultrasonic sensors, and it can supplement or even replace some LiDAR systems. As such, there is huge potential for applications in automated or autonomous driving, where vehicles need precise information on their current position and heading. This information is crucial for the vehicle such that it can plan and adjust its route both safely and efficiently.
To achieve these goals, the vSLAM algorithm developed by IAV uses live camera images. By repeatedly detecting feature points within the images, a three-dimensional map of the surroundings is created in real time. At the same time, the vehicle determines its own position within this map. Common localization solutions are based on GPS data. If the signal quality is impaired – which often happens in underground car parks or within dense inner city environments – positioning accuracy varies. At worst, the signal is lost completely. Thanks to IAV’s vSLAM solution, however, this can be avoided. The fusion of vSLAM with existing sensor data creates a robust localization solution with a wide range of applications.
IAV offers this software solution to vehicle manufacturers in the advanced engineering and production-ready development stages. Many other fields can, however, also benefit from the technology. Manufacturers of navigation systems possess large amounts of high-resolution 3D data. Combined with vSLAM, this data permits highly precise navigation over long distances. Autonomous drones can navigate through unknown terrain using vSLAM, and to return safely back to their launch site. Would you like to learn more about the technology and its potential applications? Please get in touch with us.