This paper examines the integration of Connected, Cooperative, and Automated Mobility (CCAM) in public transportation, focusing on automated buses. It identifies key challenges such as legislative frameworks, human factors, and infrastructural conditions that impede deployment in Europe. The paper proposes a combination of standardized processes and interfaces, scalable software architecture as well as V2X communication to address these issues. In addition, it discusses projects like MINGA, OTRACE, and Country2City-Bridge (C2CBridge) that exemplify the application of these technologies. Lastly, the gradual introduction of autonomous transport technologies, such as platooning, is recommended to evaluate and test the automated capabilities as well as build public confidence before full automation.

Connectivity, whether between objects or with the Internet, has given rise to many new business ideas and associated terminologies. Especially in the automotive industry, new terms have emerged in the field of connected vehicles, which are defined multiple times in academia and industry. This is due to the lack of an ontology that establishes the connection between terminologies and a taxonomy that categorizes them clearly. Therefore, this paper closes this gap by providing an overview of paradigms, concepts, and technologies for vehicle communication, as well as potential execution platforms and concepts.Additionally, we propose use cases that deal with the connected vehicle in the commercial city bus sector. The first use case deals with the connectivity of the vehicle and the associated possibility of executing computationally intensive machine learning models in the cloud. The aim of this use case is to reduce the energy impact of air conditioning in the city bus. The second use case focuses on cooperative perception and cloud/edge connectivity to increase the efficiency of city buses in their daily operation in road traffic by implementing dynamic route optimization that can react promptly to events such as accidents. The paper closes with remarks on future research challenges in the fields of function offloading, the identification of the most suitable execution concept for vehicle functions and the selection of the most fitting variant of a function in the cloud.

The modern vehicle is increasingly becoming a part of the Internet of Things (IoT). Video or audio streaming from cloud resources are already state-of-the-art. However, vehicle connectivity opens up even more possibilities by extending the vehicle's electrical/electronic architecture (E/E architecture) to the cloud. That said, vehicle functions that are currently deployed and calculated on onboard electronic control units (ECUs) can be offloaded to the cloud. Not every function is implementable in the cloud. The implementation of motor control in the cloud seems unrealistic without a doubt. It is therefore necessary to identify the relevant requirements for a function to decide on the technical feasibility and suitability for offloading a function to the cloud. This paper proposes a stepwise process that classifies the technical feasibility and then calculates the degree of suitability based on defined quantitative and qualitative evaluation grades. The suitability is assessed with a weighted scoring system. The process is finally applied using the vehicle functions that we have found within the E/E architecture of a modern premium vehicle, and the most suitable functions for offloading are identified.

Platooning is a step toward fully automated driving. It refers to a group of vehicles driving in close formation, with the lead-vehicle controlled by a human driver and the others followed by an automated system. The technology behind platooning involves communication between vehicles via various sensors and wireless systems that allow the vehicles to drive in close proximity to each other. Knowing the relative position of the vehicles in the platoon is essential to coordinate the movements of all involved vehicles and ensure that the platoon can travel efficiently and safely in a union. The objective of this research is to present a lidar-based method for relative positioning in the context of platooning with significantly reduced modeling effort. To achieve this, we propose a lidar-based method for measuring the relative position of vehicles. For this purpose, the lead-vehicle's cluster has to be detected from lidar point clouds in each measurement. The extracted cluster is fitted into an L shape, with the short side of the L representing the rear of the vehicle and the long side of the L representing one side of the vehicle. Using this method, the relative position of each vehicle can be derived from the L-shape. The system was evaluated both in simulation and on a real-world test track. The results show that the proposed system is capable of determining the relative position of two vehicles in an urban platoon.

Vehicle electrification is picking up speed in local public transport. Battery electric buses replace diesel buses and must therefore be able to serve the routes of diesel buses on an equal footing. Reducing the energy consumption of electric buses can make a major contribution to significantly increase their driving range and their possible field of application for transport companies. Therefore, an analysis of the different consumers and influencing factors on the energy consumption was conducted on the basis of real measurement data. An overview of existing optimization approaches is given, classifying them in the different layers of component-based, system-based and cloud-based optimization.

This paper uses deep learning in the field of public transport, where it is of immense interest to know the occupation of vehicles for their coordination and scheduling. Counting people on a bus or train is very often still done manually, whereas this paper presents a modern approach with just the use of camera images and deep learning. The people counter presented in this work consists of two parts, in each of which a neural network has been optimized. In the first step, persons are recognized by means of a person detector while this information is used in the second step for counting people. For the detection of people at the bus entrance, RetinaNet was selected as the model and optimized. The output of the person detector was then used to optimize a novel architecture of a neural network. This allows determining the number of people getting on and off a bus in a video.

The development of highly automated driving functions is currently one of the key drivers for the automotive industry and research. In addition to the technical constraints in the implementation of these functions, a major challenge is the verification of functional safety. Conventional approaches aiming at statistical validation in the sense of real test drives are reaching their economic limits. On the other hand, there are simulation methods that allow a lot of freedom in test case design, but whose representativeness and relevance must be proven separately. In this paper an approach is presented that allows to generate critical concrete scenarios and test cases for automated driving functions by means of a reinforcement learning based optimization using here the example of an overtaking assistant. For this purpose, a Q-Learning approach is used that automates the parameter generation for the test cases. While pure combinatorics of the variable parameters leads to an unmanageable amount of test cases, the percentage of actually relevant critical test cases is very low. In this work we show how the share of critical and thus relevant test cases can be increased significantly by using the presented method compared to a purely combinatorial parameter variation.

In Machine Learning, feature selection is paramount to achieve accurate and correct results as well as to reduce the computational effort. As a first step, it can be used to identify erroneous features or features introducing spurious correlation. This also applies to machine learning in an engineering environment. Here, data is often given as multivariate time series datasets, which require adapted approaches for machine learning as well as preprocessing steps, like feature selection. Because of the importance of feature selection, special methods for this use case are needed. Therefore, we propose an applicable feature selection method for multivariate time series, based on a differential correlation approach. This method aims to detect erroneous features or those introducing spurious correlation through the assessment of variable relationships across multiple multivariate time series instances. For this, the correlation between the time series features is compared across all examined instances through a differential correlation approach. To show the applicability of the proposed selection method, it is evaluated on artificially generated time series datasets, including erroneous features. Additionally, the method is evaluated on a real multivariate time series dataset in context of an engineering environment.

Forward Collision Warning (FCW) is an advanced driver assistance system (ADAS) designed to prevent or reduce the severity of a collision by providing early warnings to the driver. The core algorithm of a FCW-system is based on the distance between the vehicle and the obstacle as a warning threshold. Since the system assumes a particular reaction time for all drivers, the threshold value cannot be individualized for different driver states and thus leads to false or unnecessary alarms. Therefore, this paper proposes an adaptive custom collision warning system. The system relies on camera images to collect data like age, emotion, fatigue, and attention of the driver through Deep Residual Networks and the PERCLOS-method (Percentage eye openness tracking). The information obtained is evaluated using fuzzy logic and an appropriate reaction time is derived. The improved safe-distance algorithm calculates a safety distance appropriate to the driver’s condition. The system is evaluated with the National Highway Traffic Safety Administration (NHTSA) FCW standard test through the simulation environment IPG CarMaker.

There are many different procedures to carry out a data mining project. Depending on the application, various methods have to be chosen, which is normally done by experts. For inexperienced users without machine learning experience, it is difficult to analyse their data without help.To simplify the access to data mining particularly for multivariate time series analysis, we propose an intuitive to...