Advanced Drivers Assistance Systems (short: ADAS) of SAE Level 3+ need to be recognized as major technology leap. Those systems need to bring much more performance than systems of lower automation levels. Developing ADAS of SAE Level 3+ means: Build and integrate a comprehensive system of sensors, algorithms, actuators, compute power and artificial intelligence to provide answers for the four fundamental questions of vehicle automation:
Since the beginning of the collaboration of NAVENTIK’s founders as a research group at University of Technology Chemnitz in 2011, the R&D focus was to find a technical answer for the first question of vehicle automation at SAE Level 3+: “Where Am I ?”
Why can ADAS then be on the streets today? Well, today’s ADAS in serial manufactured cars just work with landmark based positioning to compute a relative position. By utilizing front cameras and ranging sensors, the ADAS recognizes lane markings and the vehicle’s relative position in between. This is the starting point to trigger an automated driving function. Preconditions to engage the ADAS are open highway surroundings, non- complex traffic situations, the simple existence of lane markings and sufficient weather conditions (no rain, no fog, no snowfall).
For the deployment of ADAS of Level 3+ with a much broader range of safety critical driving maneuvers in more complex (urban) environments, the absolute position on a HD map will become a crucial information beside the relative position. It is clear: To deploy ADAS of Level 3+ anywhere, anytime and on mass market scale, the ADAS needs maximized accuracy, availability, robustness, functional safety and scalability. This can only be achieved by adding a sufficient absolute positioning system as the complementary technology parallel to the relative positioning.
The technology gap for the deployment of a sufficient absolute positioning system for ADAS L3+ gets obvious, if a comparison along the defined requirements for an ADAS is made. The following target system for an absolute positioning, compliant to the requirements of ADAS Level 3+, can be derived:
Carrier phase based positioning engine with advanced error modeling and rapid convergence times, interference mitigation and monitoring.
Adding inertial or vision aided navigation increases availability and continuity of the solution, seamless integrated on vehicle ECU.
Integration of precise SSR based satellite and atmospheric corrections including integrity measures.
Development according to automotive safety standards (ISO26262) to meet Automotive Safety Integrity Levels (ASIL).
Both technology components can be deployed as standalone products – independently from each other. For example, PATHFINDER emPE works as positioning engine also in combination with any third party GNSS receivers. But with both components combined, NAVENTIK provides the most comprehensive automotive positioning platform for safety critical applications – out of one hand. The derived final integration architecture for ADAS looks as follows:
“It totally makes sense. Why should I buy an RTK/INS System with a costly IMU, if my ADAS system is 5G connected and already has a good ASIL IMU deployed for chassis controls? Let’s just combine this with GNSS!”
Current release of PATHFINDER GNSS & PATHFINDER emPE available under R&D License
(Software libraries for integration on your ADAS platform or delivered on NVIDIA® Jetson NX™)
State Space Representation (SSR) correction services are the latest generation of GNSS correction services provided by new entrants and the larger legacy providers. The term state is used as the core techniques are centered on using the reference station network to model the key errors over large geographical regions and transferring them to the rover. The rover uses the incoming data to create a local model of the GNSS errors, which it applies directly to the GNSS observations. The performance of SSR correction services depends on what errors are transmitted to the rover device, which is detailed in the table below.
