Self-learning software for development of predictive driving launched by Elektrobit
Elektrobit (EB) has announced EB robinos Predictor, a software module that provides street map and topographical data for development of predictive driving systems. The new solution features a self-learning algorithm to enable the implementation of advanced driver assistance systems (ADAS), like predictive headlights, curve speed warning, and traffic-sign assistance.
The hardware-agnostic software consists of three optional components, including provider, reconstructor, and a self-learning map module. The first component, provider, is also part of the Dynamic eHorizon solution of Continental AG. EB robinos Predictor supports map and industry standards like NDS, Volkswagen PSD, BMW ADAS, and ADASIS v2. It also supports free open source OpenStreetMaps (OSM). The software uses this mapping information to generate data, which is interpreted by EB’s electronic horizon reconstructor software module into information usable by the car’s electronic control unit (ECU). This enables cars to prepare for and respond to real-world situations. Moreover, map data can be incrementally enriched over the air while the vehicle is driving, using EB’s self-learning algorithm.
“Advanced driver assistance systems not only make driving more enjoyable, but also much safer,” said Dr. Björn Giesler, Head of Driver Assistance at EB. “Financial restraints shouldn’t hold carmakers and suppliers back from offering these features to all of their customers, no matter which vehicle model they’re driving. With our open standards approach, we hope to help automakers streamline the development and production of highly automated driving systems, and get them on the fast track to automated driving.”
In 2017, EB expects to release additional, complementary software modules, including EB robinos Grid Fusion and EB robinos Path Planning, as well as an accurate positioning module. EB robinos Predictor is available now. For more information please visit https://www.elektrobit.com/.