Last modified more than a year ago

Wanted: more realism and greater connectivity

30 April 2021

As a carmaker, we are often asked how long people will have to wait for self-driving cars. However, despite the marketing-driven reports from industry newcomers leading us to believe that the age of autonomous Level 5 robotaxis is just around the corner, in reality we still have a considerable way to go before they are ready to be industrialised.

Dr Nicolai Martin, BMW Group

The reason for this? The challenges for automated driving are basically defined by two factors: 1) The speed of the vehicle: the faster it’s moving, the greater the potential severity of the accidents that need to be avoided; 2) The complexity of the traffic scenarios: the greater the number of different road users and directions of travel involved, the more challenging it is to control the situation. This is why automated vehicle operation is only possible at present when travelling either at higher speeds in less complex situations (e.g. on motorways) or at low speeds in more complex environments, as is the case with autonomous parking. With today’s Level 2 functions, the speeds in simple traffic situations are being steadily increased and ever more complex traffic situations are gradually coming into play.

The next step is the implementation of Level 3 systems. Level 3 tailback and slow-moving traffic systems for motorways offer a good balance between the two counteracting aspects mentioned above. Lawmakers in the EU – and in Germany, in particular – are already giving us all the support we need for implementation on a regulatory level.

Automated driving at BMW

At the BMW Group, our primary focus is on offering our customers thoroughly developed, high-quality functions that assist them with the task of driving and ease their workload. We currently already have around 40 driver assistance functions in our vehicles that help with driving and parking. These encompass everything from the High Beam Assistant to the rear-view camera and intelligent cruise control with longitudinal and lateral guidance. Not only have these systems been extremely well received worldwide, they have been shown to make driving approximately 30% safer[1].

One new feature (available in Germany) is the supplementary “traffic light recognition” function for Active Cruise Control with Stop&Go. This function represents a major step towards automation in urban traffic. Our traffic light recognition feature comprises two sub-functions: “reaction to red lights” – as part of the Active Cruise Control with Stop&Go functionality – and a “start-off prompt”. This means the vehicle clearly recognises the colour of the lights at simple traffic lights and will automatically come to a halt at the appropriate stop line if it is at the front of the queue and the function is active. We are the first carmaker to offer this feature. The traffic light recognition function is already so sophisticated that it is able to detect more complex sets of traffic lights in advance. If the system cannot definitively determine which light is of relevance for the driver, additional confirmation is obtained from the driver before automatically bringing the vehicle to a halt. This concept prevents frequent erroneous braking – e.g. at green lights or lights where the relevant colour cannot be clearly identified – while also maximising automation at simple sets of traffic lights.

Communication with digital infrastructure promises to be the game changer.

A logical next step in this context would be the integration of digital traffic light controls, i.e. a system that could additionally assist traffic flows. Here, as with other functions, we are pinning our hopes on the necessary modifications to infrastructure. Together with data from the cloud (“Mobility Data Marketplace”, in future “Mobility Data Room”), this could enable our assistance systems to provide even greater value and thereby bring about a sustained improvement in both road safety for everyone and traffic flows[2]. The more vehicles are using these technologies, the sooner we can achieve a positive ecological impact across road traffic and reduce emissions. The European data infrastructure initiative GAIA-X will make it possible to share the necessary data securely between different systems and allow it to be utilised. Here, the protection of customer data is of paramount importance for the BMW Group.

Cooperation and Connectivity is the key.

We view information-sharing as vital here. At the BMW Group we see ourselves as having a clear responsibility to not just call for international standards in this area, but also set them and consistently refine them. This is illustrated by the development work now getting underway on ISO technical specification 5083 “Road vehicles – Safety and cybersecurity for automated driving systems — Design, verification and validation”[3] as part of a collaboration between experts from over 50 manufacturers and suppliers around the world. BMW is the ISO project lead. This standard builds on the work carried out for ISO/TR 4804, which was published in December 2020 under the same title.

Beyond technology and standardisation, BMW is strongly involved in EU funded research projects under Horizon 2020, InterACT and L3Pilot to name the most recent and important ones. Thecooperation and coordination of the whole programme by the support project ARCADE and our contribution there shows our full commitment to being not only standardisation leader, but also contributing to set the European framework in research and regulation. The central place to look at for all this the CAD Knowledge Base and its content on data sharing.

We also warmly welcome and support the new measures such as the CCAM Platform and the Horizon Europe research programme in paving the way for a wealth of important automation functions. We believe that sharing all data not relevant to competitors within this framework will serve to keep Europe competitive in the field. Collaboration isn’t just a necessity, then, it is a wise move – and we are glad to be a part of it.

Dr Nicolai Martin

Senior Vice President Automated Driving and Driver Assistance Systems, BMW Group

[1] 2014-2018 study in the USA involving one million vehicles demonstrated that vehicles equipped with assistance systems had 30% fewer accidents than vehicles without assistance systems.

[2] 4th interim report from Germany’s National Platform Future of Mobility (NPM) – WG 3: Measures for the digitalisation of transport infrastructure

[3] This is based on the white paper “Safety First for Automated Driving” with contributions from just under 50 international experts from Japan (Toyota, Nissan, Honda, Denso), China (Baidu), South Korea (KAATA), France (PSA, Renault, Valeo), Austria (AIT, TTTech), Belgium (Toyota), the Netherlands, Italy (Nvidia), Germany (Daimler, Volkswagen, Bosch, Continental, Denso, exida, TÜV Süd, Infineon), the UK (ARM, MIRA, Univ. of Warwick) and the USA (FCA, Intel, Here, Aptiv).