Bosch und die VW-Softwaretochter CARIAD wollen bei der Entwicklung selbstfahrender Autos zusammenarbeiten. Dazu wurde dazu eine umfassende Partnerschaft vereinbart.
The companies want to make partially and highly automated driving suitable for volume production, and thus available to the broad mass of consumers. For the vehicles sold under the Volkswagen Group brands, the alliance aims to make functions available that will allow drivers to temporarily take their hands off the steering wheel. More specifically, these functions are Level 2 hands-free systems for urban, extra-urban, and freeway driving, as well as a system that takes over all driving functions on the freeway (SAE Level 3). The first of these functions are to be installed in 2023.
For partially and highly automated driving, the two companies will jointly develop a state-of-the-art, standardized software platform. The aim is to use this platform in all privately used vehicle classes sold under Volkswagen Group brands – and thus in one of the world’s biggest vehicle fleets. It will also be possible to integrate all the component parts developed by the alliance in other automakers’ vehicles and ecosystems.
At various locations belonging to the two companies, and especially in Stuttgart and Ingolstadt, associates from the Bosch Cross-Domain Computing Solutions division and CARIAD will work together to develop partially and highly automated driving functions. Working in mixed, agile teams, they will be part of a global network. At peak times, it is expected that the various modules required by the project – from middleware to individual applications – will occupy more than 1,000 experts from the two companies. The two companies have already started recruiting fresh experts to work in the alliance as well.
The focus of the project’s work will be data-driven software development on the basis of information from 360-degree surround sensing. For this purpose, a highly innovative development environment will be created for the recording, evaluation, and processing of data. The environment will also make use of AI methods. The idea behind this is a simple one. The more extensive the pool of data from real road traffic, the more robust and natural the design of the partially and highly automated driving functions can be. This applies, for example, to additional layers for high-resolution maps for the localization and lateral and longitudinal guidance of vehicles. The alliance will also be working on these layers. Moreover, it applies equally to everyday driving situations and to what are known as “corner cases” – the seldom occurring road-traffic incidents that are especially tricky for a system to resolve.