Search

The progress towards fully autonomous mobility is significantly impacted by the integration of evolving technologies in connected and automated mobility (CAM). Connected and automated vehicles (CAVs) have the potential to revolutionize our transportation system by improving efficiency, safety, and environmental sustainability. Automated shuttles and public buses, smart traffic signals, intelligent passenger cars, and smart roundabouts are just a few examples of technologies that are being planned and actively researched for integration into transportation systems. Sensors are essential in making this possible. This article provides a structured overview of research on the selection and positioning of sensors on- and off-vehicle to achieve cooperative, connected, and automated mobility. The general integration and usage of sensors in vehicles and infrastructure is described, a detailed taxonomy of the examined research is provided, and future research directions are presented, involving solutions for quantification of sensor performance and addressing current trends. The findings of this article also highlight numerous challenges in creating a universal framework, the lack of application of novel machine learning methods, and the complexity of modeling multi-sensor settings.

Power systems are increasingly confronted with operational uncertainties. These stem from the integration of wind and solar energy sources with inherently stochastic generation behavior. The ensuing risks include power plant curtailments or grid imbalances. The deterministic approaches that currently underpin most planning have proven insufficient to manage these risks. This paper presents an alternative in the form of a data-driven probabilistic approach with direct relevance for transmission system operators. The models present a novel methodology that is independent of the common boundary conditions, e.g. case-specific models. First, we compare several data-driven algorithms and assign the forecasting task to the best-performing one. Second, the resulting forecast serves as an input for three optimal power flow (OPF) problems we tailor to the German power system. These problems minimize energy import volumes, energy import costs, and overall power losses. Third, based on the OPF results, we perform a risk assessment for operational instability, power loss, financial losses, and renewable energy waste. The results show that neural networks slightly outperform traditional machine learning algorithms in forecasting accuracy. However, linear-quadratic regulators remain attractive for their simplicity-performance ratio. Our probabilistic OPF approach can reduce power losses and identify frequency and line loading irregularities that deterministic methods do not. The data-driven approach we propose is superior to existing approaches in terms of its performance, usability, and applicability to complex power systems.

The efficient processing of information plays a critical role in data-driven domains. Data acquisition pipelines, which act as the interface between data collection and subsequent processing, are central to this. Traditional software engineering methods form the foundation for the development of robust data acquisition pipelines, but agent-based systems hold great potential for realizing such pipelines, and this study presents an innovative architecture that uses agent-based components. This modular approach makes it possible to use specialized agents for individual tasks and to increase the effectiveness of the pipelines. The applicability and effectiveness of this architecture are demonstrated using a system for autonomous driving that collects and processes data from edge computers along roads. Our results are on GitHub (https://github.com/BenjaminAcar/Agent-based-Data-Acquisition-Pipeline).

Abstract The future of driving is autonomous. It requires a comprehensive stack of embedded software components, enabled by open-source and proprietary platforms at different abstraction layers, and then operating within a larger ecosystem. Autonomous driving demands connectivity, cooperation and automation to form the cornerstone of autonomous mobility solutions. Platform economy principles have revolutionized the way we produce, deliver and consume products and services worldwide. More and more businesses in the field of mobility and transport appear to implement transaction, innovation, and integration platforms as core enablers for Mobility-as-a-Service and transport applications. Artificial intelligence approaches, especially those dealing with distributed systems, enable new mobility solutions, such as autonomous driving. This paper contributes to understanding the intertwining role between distributed artificial intelligence, autonomous mobility and the resulting platform ecosystem. A systematic literature review is applied, in order to identify the intersection between those aspects. Furthermore, the research project BeIntelli is considered as a hands-on application of our findings. Taking into account our analysis and the aforementioned research project, we pose a blueprint architecture for autonomous mobility. This architecture is the subject of further research. Our conclusions facilitate the development and implementation of future urban transportation systems and resulting mobility ecosystems in practice.

This paper presents a comprehensive study on generating subjective inquiries for news media posts to empower public engagement with trending media topics. While previous studies primarily focused on factual and objective questions with explicit or implicit answers in the text, this research concentrates on automatically generating subjective questions to directly elicit personal preference from individuals based on a given text. The research methodology involves the application of fine-tuning techniques across multiple iterations of flan-T5 and GPT3 architectures for the task of Seq2Seq generation. This approach is meticulously evaluated using a custom dataset comprising 40,000 news articles along with human-generated questions. Furthermore, a comparative analysis is conducted using zero-shot prompting via GPT-3.5, juxtaposing the performance of fine-tuned models against a significantly larger language model. The study grapples with the inherent challenges tied to evaluating opinion-based question generation due to its subjective nature and the inherent uncertainty in determining answers. A thorough investigation and comparison of two transformer architectures are undertaken utilizing conventional lexical overlap metrics such as BLEU, ROUGE, and METEOR, alongside semantic similarity metrics encompassing BERTScore, BLEURT, and answerability scores such as QAScore, and RQUGE. The findings underscore the marked superiority of the flan-T5 model over GPT3, substantiated not only by quantitative metrics but also through human evaluations. The paper introduces Opinerium based on the open-source flan-T5-Large model, identified as the pacesetter in generating subjective questions. Additionally, we assessed all aforementioned metrics thoroughly by investigating the pairwise Spearman correlation analysis to identify robust metrics.

While multi-agent frameworks can provide many advanced features, they often suffer from not being able to seamlessly interact with the outside world, e.g., with web-services or other multi-agent frameworks. This may be one factor that hinders a broader application of multi-agent systems in production systems. A possible solution to this problem is the combination of multi-agent systems with the concepts of micro-services and containerization, providing language-agnostic open interfaces, as well as encapsulation and modularity. In this paper, we propose an API and reference implementation that can be employed by multi-agent systems based on different languages and frameworks. Each agent component is encapsulated in a container and is accessed through its parent runtime platform, which takes care of aspects such as authentication, input validation, monitoring and other infrastructure tasks. Multiple runtime platforms can then be connected to form systems of distributed, heterogeneous multi-agent societies.

The application of containerization technology has seen a significant increase in popularity in recent years, both in the business and scientific sectors. In particular, the ability to create portable applications that can be deployed on different machines has become a valuable asset. Autonomous driving has embraced this technology, as it offers a wide range of potential applications, including the operation of autonomous vehicles and the digitization of infrastructure for the development of Cooperative, Connected, and Automated Mobility (CCAM) services. This paper provides a comprehensive analysis of containerization in autonomous driving, emphasizing its application, utility, benefits, and limitations.

The radical increase in the popularity of electric vehicles (EVs) has in turn increased the number of associated problems. Long waiting times at charging stations are a major barrier to the widespread adoption of EVs. Therefore, battery swapping stations (BSSs) are an efficient solution that considers short waiting times and healthy recharging cycles for battery systems. Moreover, swapping stations have emerged as a great opportunity not only for EVs, but also for power systems, with regulation services that can be provided to the grid particularly for small networks, such as microgrid (MG) systems. In this study, the optimum location and size that maximize the revenue of a swap station in an MG system are investigated. To the best of our knowledge, this study is first to solve the placing and sizing problem in the MG from the perspective of a BSS. The results indicate that bus 23 is the BSS’s optimal location and is crucial for maximizing revenue and addressing issues like the provision of ancillary services in microgrid system. Finally, the swap demand profile of the station serving electric bus public transportation system was obtained using an analytical model based on public transportation data collected in Berlin, Germany.

Multicast communication over wireless networks has many potential applications, such as real-time audiovisual content distribution or digital signage systems with multiple remote terminals. However, today’s common 802.11 networks cannot fully support such applications at the link layer due to the technical challenges in achieving both high transmission rate and high reliability for multicast data transfers. Those challenges primarily emerge from the inapplicability of immediate acknowledgment mechanisms of unicast transmissions to the multicast case and the need for finding a common transmission rate for all receivers with diverse channel conditions. The research literature in this domain mainly addresses the problem at the higher layers of the protocol stack. In this article, we present a novel approach that combines wireless link rate selection with an adaptive packet-level Forward Error Correction (FEC) mechanism in order to achieve high-rate and highly reliable multicast in 802.11 wireless networks. The integration of FEC into the 802.11 MAC layer allows direct interaction between the transmission rate and the coding scheme. As a result, potential packet losses caused by higher link rates can be compensated by the adjustable redundancy provided by FEC. In combination with an aggregated receiver feedback mechanism, this yields improved transmission efficiency and reliability for wireless multicast. We investigate this approach in a simulation environment under a realistic wireless channel model in various application scenarios with up to 50 receivers. The results represent significant performance improvements, in terms of throughput, channel utilization, and packet loss, over the state-of-the-art methods for reliable wireless multicast.

The recent developments in electric vehicles (EVs) causes several issues that have not been satisfactorily addressed. One of the foremost problems is the charging–discharging processes of EV batteries with diverse characteristics. Although a charging station is the first choice in this regard, a battery swap station (BSS) is also a suitable alternative solution as it eliminates long waiting periods and battery degradation due to fast charging. BSS has the capability to ensure prompt and efficient service for electric vehicles. Since BSS has a large number of battery systems, optimum planning of the charging–discharging operations of the batteries is critical for both BSS and the grid. This study presents an optimal charging–discharging schedule for multiple BSSs based on the swap demand of privately owned EVs and electric bus (EB) public transportation system. In addition, BSSs reinforce the power grid by providing ancillary services such as peak shaving and valley filling with demand response programs. In order to increase the flexibility of the operation, the mobile swapping station (MSS) concept, an innovative and dynamic service, is introduced to the literature and added to the model. The results indicate that BSS is an essential agent in the ancillary services market and the MSS concept is a yielding solution for both BSSs and power networks. Last, the data utilized in the study for swap demand calculation and power grid analysis are real-world data from Berlin, Germany.

The structural mass of a building provides inherent thermal storage capability. Through sector coupling, the building mass can provide additional flexibility to the electric power system, using, for instance, combined heat and power plants or power-to-heat. In this work, a mathematical model of building inertia thermal energy storage (BITES) for integration into optimized smart grid control is introduced. It is shown how necessary model parameters can be obtained using generalized additive modeling (GAM) based on measurable building data. For this purpose, it is demonstrated that the ceiling surface temperature can serve as a proxy for the current state of energy. This allows for real-world implementation using only temperature sensors as additionally required hardware. Compared with linear modeling, GAM enable improved modeling of the nonlinear characteristics and interactions of external factors influencing the storage operation. Two case studies demonstrate the potential of using building storage as part of a virtual power plant (VPP) for optimized smart grid control. In the first case study, BITES is compared with conventionally used hot water tanks, revealing economic benefits for both the VPP and building operator. The second case study investigates the potential for savings in CO2 emission and grid connection capacity. It shows similar benefits when using BITES compared to using battery storage, without the need for hardware investment. Given the ubiquity of buildings and the recent advances in building control systems, BITES offers great potential as an additional source of flexibility to the low-carbon energy systems of the future.

Power system frequency plays a pivotal role in ensuring the security, adequacy, and integrity of a power system. While some frequency response services are automatically delivered to maintain the frequency within the stipulated limits, certain cases may require that system operators (SOs) manually intervene-against the clock-to take the necessary preventive or corrective actions. As such, SOs can be greatly aided by practical tools that afford them greater temporal leeway. To this end, we propose a methodology to forecast the power system frequency in the subsequent minute. We perform an extensive analysis so as to identify the factors that influence power system frequency. By effectively exploiting the identified factors, we develop a forecasting methodology that harnesses the long short-term memory model. We demonstrate the effectiveness of the proposed methodology on Great Britain transmission system frequency data using comparative assessments with selected benchmarks based on various evaluation metrics.

This paper addresses the problem of predicting time series data using the autoregressive integrated moving average (ARIMA) model in an online manner. Existing algorithms require model selection, which is time consuming and unsuitable for the setting of online learning. Using adaptive online learning techniques, we develop algorithms for fitting ARIMA models without hyperparameters. The regret analysis and experiments on both synthetic and real-world datasets show that the performance of the proposed algorithms can be guaranteed in both theory and practice.

The Internet of Things and Services (IoTS) has encouraged the development of service provisioning systems in respect to Smart City topics. Most of them are operated as heterogeneous systems which limits end customers’ access and contradicts with IoTS principles. In this paper, we discuss and develop a reference model of an interconnected service marketplace ecosystem. The prototypical implementation incorporates findings from an empirical study and lessons learned from research projects. The elaborated ecosystem enables service request roaming between different parties across system boundaries. The paper presents a feasible centralized architecture, introduces involved parties and parts of a developed message protocol. Why a contracting mechanism is indispensable for request roaming is also outlined. The model’s feasibility is demonstrated by means of a current electric mobility use case: providing access to foreign charging infrastructure without multiple registrations. This work contributes to simplify the data exchange between service platforms to improve Smart City solutions and to support travelers with intelligent mobility applications.

Rings are widely accepted wearables for gesture interaction. However, most rings can sense only the motion of one finger or the whole hand. We present PeriSense, a ring-shaped interaction device enabling multi-finger gesture interaction. Gestures of the finger wearing ring and its adjacent fingers are sensed by measuring capacitive proximity between electrodes and human skin. Our main contribution is the determination of PeriSense’s interaction space involving the evaluation of capabilities and limitations. We introduce a prototype named PeriSense, analyze the sensor resolution at different distances, and evaluate finger gestures and unistroke gestures based on gesture sets allowing the determination of the strengths and limitations. We show that PeriSense is able to sense the change of conductive objects reliably up to 2.5 cm. Furthermore, we show that this capability enables different interaction techniques such as multi-finger gesture recognition or two-handed unistroke input.

This paper describes an algorithm for semantic decomposition. For that we surveys languages used to enrich contextual information with semantic descriptions. Such descriptions can be e.g. applied to enable reasoning when collecting vast amounts of information. In particular, we focus on the elements of the languages that make up their semantic. To do so, we compare the expressiveness of the well-known languages OWL, PDDL and MOF with a theory from linguistic called the Natural Semantic Metalanguage. We then analyze how the semantic of the language is build up and describe how semantic decomposition based on the semantic primes can be used for a so called mental lexicon. This mental lexicon can be used to reason upon semantic service description in the research domain of service match making.

The number of unmanned aerial system (UAS) applications for supporting rescue forces is growing in recent years. Nevertheless, the analysis of sensed information and control of unmanned aerial vehicle (UAV) creates an enormous psychological and emotional load for the involved humans especially in critical and hectic situations. The introduced research project EffFeu (Efficient Operation of Unmanned Aerial Vehicle for Industrial Firefighters) especially focuses on a holistic integration of UAS in the daily work of industrial firefighters. This is done by enabling autonomous mission-guided control on top of the presented overall system architecture, goal-oriented high-level task control, comprehensive localisation process combining several approaches to enable the transition from and to GNSS-supported and GNSS-denied environments, as well as a deep-learning based object recognition of relevant entities. This work describes the concepts, current stage, and first evaluation results of the research project.

Providers of next-generation ubiquitous services are facing increasing competition. In order to attract and satisfy customers, these services must offer added value e.g. by delivering personalized information in a seamless and multi-modal way. Personalized services, however, bring about additional requirements related to privacy and security, which have to be addressed if the services are to become widely accepted. Keywords: personalization, privacy, security, multi-agent systems, contextawareness.