Up to now, self-driving cars and trucks only function when maps are extremely precise and under moderate weather conditions which result in moderate road conditions. With "MapLite", researchers at MIT have now developed a framework that allows vehicles to improvise and find their way even under difficult conditions and without 3D maps. This has been reported by the online magazine "Engadget" . The vehicles determine their location by GPS and calculate both the way to the actual final destination and to a target location within the range that its sensors can detect. By means of these sensors, the system also recognizes the sides of the road. MapLite is not currently suited for all situations; mountain roads and other roads with extreme height differences, for example, still cause problems.
According to experts, autonomous vehicles will be dominating the streets in just a few years. For example, Renault recently launched its study "EZ-GO" at the Geneva Motor Show: an autonomous electric taxi that is designed to carry up to six people safely - but only at a top speed of 50 km/h and only in an inner-city area. The KI platform "ZF ProAI" , which the automotive supplier ZF Friedrichshafen has developed in cooperation with Nvidia, is set to reach series production this year. This platform will initially enable autonomous driving only in well-developed areas.