Refine
Year of publication
Document Type
- Conference Proceeding (25) (remove)
Language
- English (25)
Is part of the Bibliography
- no (25)
Keywords
- 02 (4)
- 3GPP (1)
- 5G (3)
- 5G mobile communication (1)
- 5G use case (1)
- 5G-Technology (1)
- Agriculture (1)
- Big Data (1)
- BigPro (1)
- CPS (4)
Institute
- Informationsmanagement (25) (remove)
Feeding the growing world population is a scientific and economic challenge. The target variables to be optimised are the yield that can be produced on a given area and the reduction of the resources used for this purpose. High-wage countries are faced with the problem that the use of personnel is a significant cost driver. Developing countries, on the other hand, usually operate on much smaller field sizes, so that the work in the field is still strongly characterised by manual labour. One solution to meet these challenges is the use of smaller autonomous harvesting robots. These can be networked into a swarm of machines to work even larger fields. The networking of autonomous agricultural machines is a key use case for rural 5G networks. 5G technology can offer many advantages over older mobile communications standards and therefore make use cases more efficient or enable new ones. Various use cases are also conceivable in the field of agriculture, yet it is unclear how 5G networks can and must be specified for this purpose. In this paper, using the example of 5G-connected harvesters powered by swarm robotics, we present the challenges that have arisen and the specification that has been developed.
For most industries, Artificial Intelligence (AI) holds substantial potentials. In the last decades, the extent of data created worldwide is exponentially increasing, and this trend is likely to continue. However, despite the prospects, many companies are not yet using AI at all or not generating added value. Often, an AI project does not exceed its pilot phase and is not scaled up. The problems to create value from AI applications in companies are manifold, especially since AI itself is diverse and there is no ‘one size fits all’ approach. One often stated obstacle, why many AI projects fail, is a missing AI strategy. This leads to isolated solutions, which do not consider synergies, scalability and seldom result in added value for the company. To create a company-specific AI strategy with a top-down approach, a generic but holistic framework is needed. This paper proposes a strategic AI procedure model that enables companies to define a specific AI strategy for successfully implementing AI solutions. In addition, we demonstrate in this paper how we apply the introduced strategic AI procedure model on an AI-based flexible monitoring and regulation system for power distribution grid operators in the context of an ongoing research project.
Crises are becoming more and more frequent. Whether natural disasters, economic crises, political events, or a pandemic - the right action mitigates the impact. The PAIRS project plans to minimize the surprise effect of these and to recommend appropriate actions based on data using artificial intelligence (AI). This paper conceptualizes a cascading model based on scenario technique, which acts as the basic approach in the project. The long-term discipline of scenario technique is integrated into the discipline of crisis management to enable short-term and continuous crises management in an automated manner. For this purpose, a practical crisis definition is given and interpreted as a process. Then, a cascading model is derived in which crises are continuously thought through using the scenario technique and three types of observations are classified: Incidents, disturbances, and crises. The presented model is exemplified within a non-technical application of a use case in the context of humanitarian logistics and the COVID-19 pandemic. Furthermore, first technical insights from the field of AI are given in the form of a semantic description composing a knowledge graph. In summary, a conceptual model is presented to enable situation-based crisis management with automated scenario generation by combining the two disciplines of crisis management with scenario technique.
In road haulage, transports are interrupted by truck drivers to comply with driving and rest times. On long-distance routes, these interruptions lead to a considerable increase in transport time. Transport interruption can be avoided by so-called relay traffic: a vehicle (e. g. semi-trailer) is handed over to a rested driver at the end of the driving time. This type of transport requires a certain company size. In Germany, however, transport companies have 11 employees on average. Intra-company relay traffic is therefore not economically viable for most transport companies. To organize an intermodal transport across forwarding companies, long-distance routes need to be split into partial routes to divide them between freight forwarders and carriers. This paper presents a data concept for an algorithm to find the best possible route sections along a previously defined start and endpoint. The developed data concept includes order-specific data, forwarder-specific data, real-time traffic data, geographical data as well as data from freight forwarding software and telematics to be the basis for the route sectioning algorithm. In this paper, different data sources, external services and logistic systems are analyzed and evaluated. It is shown which data is needed and what the best ways are to select and derive this data from the different data sources.
It is crucial today that economies harness renewable energies and integrate them into the existing grid. Conventionally, energy has been generated based on forecasts of peak and low demands. Renewable energy can neither be produced on demand nor stored efficiently. Thus, the aim of this paper is to evaluate Deep Learning-based forecasts of energy consumption to align energy consumption with renewable energy production. Using a dataset from a use-case related to landfill leachate management, multiple prediction models were used to forecast energy demand.The results were validated based on the same dataset from the recycling industry. Shallow models showed the lowest Mean Absolute Percentage Error (MAPE), significantly outperforming a persistence baseline for both, long-term (30 days), mid-term (7 days) and short-term (1 day) forecasts. A potential decrease of up to 23% in peak energy demand was found that could lead to a reduction of 3,091 kg in CO2-emissions per year. Our approach requires low finanacial investments for energy-management hardware, making it suitable for usage in Small and Medium sized Enterprises (SMEs).