Sie befinden sich hier:

ZIM Digitale Straße

High-performance Sensing and Cloud-based Realtime Data Processing for a Digital Road in Urban and Longdistance Traffic

Economic requirements regarding logistics, for example, include a more efficient utilisation of traffic routes by vehicles that act both intelligently and autonomously. However, smart cars and trucks are only one step towards a modern traffic system. A digitised transportation infrastructure, the Digital Road, is also needed to satisfy future road traffic requirements.

Within the project “High-performance Sensing and Cloud-based Real-time Data Processing for a Digital Road in Urban and Longdistance Traffic”, the Dortmund University of Applied Sciences and Arts, the Technical University of Dortmund and the industrial partner Wilhelm Schröder GmbH are working together on tasks related to the implementation of the Digital Road. The project partners focus on comprehensive capturing of traffic data in real-time. This results in various application scenarios that include the immediate detection and notification of vehicles moving in the wrong direction, demand driven traffic flow control in urban areas and statistical analysis of traffic and parking data. The prototype developed by the project partners integrates the necessary sensor technology in delineator posts on the roadside, which supercedes the costly installation of induction loops for vehicle detection and classification.

The FH Dortmund project team implements the Smart Data Platform, i.e. the software responsible for acquiring, processing and analysing the captured sensor data. Fig. 1 depicts the process to be implemented by the project partners. It comprises the following steps:

• Step 1: Capturing of Raw Data 
Leveraging radio tomography techniques the sensors in the delineator posts measure temporal data of passing vehicles.

• Step 2: Detection of Vehicle Characteristics 
The utilization of a variety of pattern recognition methods allows the identification of lane, direction, speed and type (truck, car etc.) of a vehicle. Based on the captured data the detection algorithms are iteratively improved.

• Step 3: Sensor Communication 
The sensor dataset enriched with the detected vehicle attributes is transmitted to the Smart Data Platform either wirelessly or wire-based.

• Step 4: Interface Declaration 
The platform offers the possibility for flexible provisioning of input and output data interfaces. The project team employs the Model-Driven Software Engineering paradigm to develop domain-specific languages, which enable the integration of new communication channels at runtime. Requirements of the data format and communication protocol of the respective interface are considered, with the result that new sensor types or adapted transmission formats may be incorporated in the platform without downtime.

• Step 5: Data Processing 
Prioritized Queueing enables the platform to analyze critical data prior to those of lower importance. Hence, the communication of time sensitive traffic events like wrong-way driver detection is transmitted before processing information that doesn’t require immediate action, e.g. statistical data. As with the interfaces, the computation processes for data aggregation are runtime adaptable. Thus, methods for calculating the number of passing vehicles or the number of vehicles in a certain parking zone at any given time, can be flexibly integrated in the platform and supplied via the defined interfaces.