ZEBBRA: Event based identification and assessment of bridge conditions based on radar sensors in combination with intelligent algorithms

The safety and availability of critical transport infrastructures like bridges is highly dependent on the monitoring of these infrastructures. Damages of the building structure are difficult to detect in an early stage, which is why the actual condition of a bridge often remains uncertain. Traditionally, directly contacting sensors such as strain gauges are used to measure the structures behaviour under static and dynamic loads. These sensors require time-intensive and complex installation which decreases monitoring repetition rates. In recent years, remote sensing techniques such as ground based interferometric radar (GBR) have shown great potential as alternative monitoring techniques.


The objective of the joint research project ZEBBRA is the development of a non-invasive, mobile and innovative measurement and method approach to detect and analyse the condition of bridges during operation combined with an evaluation of the bridges' condition. The ZEBBRA project is funded within the scope "Forschung für die zivile Sicherheit 2012 bis 2017" in the specific topic civil security and infrastructure.

Within the project part of the Institute of Photogrammetry and Remote Sensing (IPF) a monitoring approach for bridges based on GBR is developed. The objective is to detect changes or damages of the bridge structure. In contrast to traditional sensors, the GBR is capable of remotely measuring the displacement of several bridge points at the same time (see figure 1). As it reaches a sampling frequency of up to 200 Hz, the vibration of the bridge due to vehicles passing over it can be observed (see figure 2). The GBR-based monitoring approach is then evaluated and compared to directly contacting sensor data.

Project Status

In order to extract information about the bridge behaviour from its vibration, the GBR measurements need to pass through several processing steps. Vertical displacements of the bridge, for example, are useful information when analysing the changes of the bridge behaviour.

At first, disturbances of external sources such as vehicles passing through the signal are removed. Afterwards, several corrections are applied. This includes among others the correction of dynamics of atmospheric parameters such as temperature and relative humidity as well as a projection of the line of sight measurements to a common coordinate system. The results of these processing steps are long continuous time series. Relevant information is then extracted from these time series for further analysis. This is accomplished with traditional signal processing and new machine learning algorithms.


We aim to further improve the accuracy and reliability of the radar processing steps, thus making the measurement approach applicable for a wide range of situations at different types of bridges. A next step is to implement an automatic extraction of the vibration parameters for comparison to the model.

The increasing load on transport infrastructure by passenger cars and commercial vehicles as well as ageing of this infrastructure cause use restrictions and economic damages due to congestions and diverted traffic. The German highway network consists of about 40.000 bridges which are an important part of the critical transport infrastructure. Damage of the building structure is difficult to detect in an early stage, which is why the actual condition of a bridge often remains uncertain.

Further information:
KIT press information



Figure 1: Schema of the GBR-based monitoring approach at a bridge

Figure 2: Exemplary vertical displacement of a selected bridge point extracted from the GBR data during vehicle crossings