Content - Session V: AI Applications in Mobility and Autonomous Driving

AI Applications in Mobility & Autonomous Driving

Session abstract

AI plays a key role for Autonomous Driving, including topics such as intelligent navigation systems, traffic regulation, next-generation delivery services, autonomous taxi services, or on-demand public transport. Market research by a major consultancy company predicts that consumer acceptance of these applications will be comparatively high. Here, consumers associate AI solutions mostly with more convenience and comfort in transport, improved safety and positive environmental effects.

Speakers in this session will discuss various technical and social aspects of applying AI in these areas. These range from the connection between AI technical advances and reform efforts in national traffic regulation (Dr. Tokuda/NICT), to the use of learned models of the world for generating synthetic data for the training and validation of autonomous vehicles (Prof. Dr. Slusallek/DFKI) and the introduction of AI-based delivery and dispatch services linked to online mobility service platforms (Mr. Yamashita/DeNa). Additional coverage focuses on intelligent transport systems for automotive as well as railway systems (Prof. Dr. Köster/DLR) and solutions for predictive maintenance (Dr. Nonne).

Evolution of IoT Services & AI-enabled Connected Cars
Chair: Dr. Hideyuki Tokuda

The convergence of IoT and AI is accelerating in many business domains. Many enablers such as wireless sensors, LPWA, and smart IoT devices are allowing us to transform connected products as connected services easily. Similarly, many AI and machine learning tools are available for supporting human activities. This type of digital transformation not only creates new values for our society but also often increases risks in daily life. In this talk, we will address the evolution of IoT services and issues in AI-enabled connected cars. We will also discuss cyber and physical attacks over connected cars as well as social/policy issues for accepting autonomous cars in our society.

Understanding the World with AI: Training & Validating Smart Machines Using Synthetic Data
Chair: Prof. Dr. Philipp Slusallek

The world around us is highly complex but Autonomous Systems must be able to reliably make accurate decisions that in many cases may even affect human lives. With Digital Reality we propose an approach where instead of only relying on real data, it learns models of the real world and uses synthetic sensor data generated via simulations, for the training and -- even more importantly -- the validation of Autonomous Systems. This is extended by a continuous process of validating the models against the real world for improving and adapting them to a changing environment.

DeNA's Challenges to Realize Future Mobility Services
Atsushi Yamashita

Recently, Japan has faced higher numbers of those vulnerable in transportation and a logistics crisis in conjunction with its aging society and manpower shortage. We believe self-driving mobility services will be effective in addressing these issues. We think "Artificial Intelligence" is the key technology for self-driving mobility services. So, we will introduce DeNA's Artificial Intelligence capability and the utilization for mobility services.

Integrated Uses of real-live Data & synthetic Data for the Development of AI-based Driving Functions
Prof. Dr. Frank Koester

Since 2014, the Institute of Transportation Systems operates the Application Platform for Intelligent Mobility (AIM). AIM allows the use of an entire city for research on automated and connected driving and so on. For this purpose, a wide range of databases, models, simulations and simulators, dedicated test tracks as well as real urban areas within Brunswick can be used to support data driven research. Currently, integrated uses of real-live data and synthetic data for AI-based vehicle functions are of high interest. The talk gives an overview on our research infrastructure, selected research activities, and future plans to extend the existing infrastructure of AIM (Test Field Lower Saxony).

AI for Railways: A Strategic Challenge for Mobility as a Service & for Industrial Operations
Dr. Héloïse Nonne

Mass transit by rail presents a real challenge in terms of operations and service quality, in particular the running of trains as part of an industrial process that occurs in an open environment with unknown and unpredictable variables. AI can help by improving and securing processes and services. Using examples showing the industrial perspective and the customer service perspective, both of which are intertwined, I will illustrate the role of AI and how data exchanges can be leveraged. I will also explain how business, open-environment and security constraints translate to AI algorithms and how the various challenges can be addressed.