Automated Drivingin the City
Automated vehicles and intelligent traffic are key components of the urban traffic of the future. Automated vehicles aim to ensure that the driver and passengers enjoy a reliable, safe and comfortable journey in all situations. The interaction between the vehicle and driver, as well as between the vehicle and pedestrians or cyclists at intersections and roundabouts, for example, is of decisive importance to ensure a journey remains safe – and necessitates a solid situational understanding. @CITY has created the foundations for safe, stress-free, efficient and comfortable driving in the city based on innovative, automated driving functions – which not only bring benefits for the driver, but for all road users.
Sensing and interpreting the environment presents a major challenge in the city. Due to the complexity of urban public roads, much more comprehensive detection systems are necessary than they are in clearly structured environments such as motorways – and must be able to overcome uncertainties and ambiguities. Because environmental sensing based exclusively on sensor systems does not currently seem to be a realistic way to achieve an acceptable traffic flow, a priori knowledge in the form of exemplary empirical knowledge and digital maps can and should be used as data sources as well.
To do this, the automated vehicle must be able to precisely pinpoint its location in relation to a highly detailed map, and check it for plausibility before using it. The information provided by the
environmental sensor systems and the data in the digital map must be combined. Only this combination, together with an extensive pool of data on road rules and behaviour, will make anticipatory situational understanding possible – and this is exactly what provides the basis for all automated driving functions for the city. One of the project’s aims was to examine whether components of this data pool can, among other things, be learned by means of machine learning processes based on observations, as well.
Using highly effective and dependable algorithms for situational understanding and precise localisation of the vehicle and an optimised configuration of automation functions as a basis, new automated driving functions for the city were developed,
implemented as prototypes and evaluated for the very first time. The main focus was directed at automated driving through complex urban junctions, passing bottlenecks with or without encountering oncoming traffic on urban thoroughfares, and the interaction with vulnerable road users such as pedestrians and cyclists. Different styles of driving behaviour, interfaces, and interactions within the triad formed by the driver, automated vehicle, and outside road users were also analysed within the project. Another distinguishing feature that makes @CITY unique was that it factors in public transport. Automated city buses will be an inherent feature of inner-city scenarios in future.
