HORIS - High-resolution radar systems for infrastructure

At a glance

If automated and autonomous vehicles are to operate efficiently and without safety concerns in the future, there is a need to put in place intelligent infrastructure with connectivity to traffic flows. A key function of such infrastructure is the capability to detect all road users, especially individual pedestrians. Though object classification technology already exists for cars, trucks, motorcycles, cyclists, and pedestrian groups, distinguishing individuals from a group of people presents a huge challenge, especially when it comes to validating the sensor technology used. This is the prime use case for HORIS – a joint Fraunhofer IIS, Fraunhofer FHR and Fraunhofer IVI project developing technology that goes beyond the state of the art.

To demonstrate the technology, a typical traffic situation was chosen: a street with two bus stops some distance apart, and road users including pedestrians and vehicles. Several radar sensors provide data, which, when fused, enables the early detection of movement if one or more people break away from the group waiting at the bus stop and spontaneously cross the road. The raw data can also be used to detect biosignals, such as pulse and breathing, which offers a way to distinguish individuals from one another and predict their behavior. The idea is to transmit a hazard map based on this information to moving traffic so emergency braking or other maneuvers can be calculated.

Measuring service

Privacy warning

With the click on the play button an external video from www.youtube.com is loaded and started. Your data is possible transferred and stored to third party. Do not start the video if you disagree. Find more about the youtube privacy statement under the following link: https://policies.google.com/privacy

Measurement systems used:

  • Qualisys motion capture system with 28 motion capture cameras
  • Two additional video cameras
  • Biopac® BioNomadix Wireless RSP with ECG Transmitter

Recreated scenarios:

  • One person breaks away from the group and crosses the street
  • The bus is at the bus stop and one person crosses the street behind it
  • Several people are waiting at the bus stop and one person crosses the street

As its primary role in the HORIS project, Fraunhofer IIS in Nuremberg carried out the measurement campaigns at the L.I.N.K. Test and Application Center, which provides a unique laboratory space for recording data from realistic test scenarios. The measurements for the application scenario were not taken at a real bus stop, but instead using realistic recreations. For these, Fraunhofer IIS used the L.I.N.K. Mobil, its mobile test bed, as a substitute for a public bus and recreated a bus stop using whiteboards. Both provide a very true-to-life environment for the radars.

 

Various real-life scenarios in dynamic road traffic were then presented using the latest in measurement technology. This setup enabled up to ten people to be tracked; by applying skeleton tracking, the team was even able to determine the position of their arms, hands, knees, legs, etc. Our measurement campaigns went as far as to determine people’s vital signs, such as their pulse or breathing, thus enabling the detection of individuals within a group. The reference data was used to train and validate machine learning and signal processing algorithms. On that basis, it was subsequently possible to predict the behavior of the pedestrians.

 

What we offer

© Fraunhofer IIS/Friesen

Measuring Service

The Fraunhofer IIS Machine Learning and Validation group is part of the Positioning and Networks business unit and has been conducting a wide variety of professional measurement campaigns for many years. Specifically, the L.I.N.K. Test and Application Center, which is regularly used for extensive campaigns, provides a unique laboratory space for recording data from realistic test scenarios. With our L.I.N.K. Mobil mobile test bed, we can also conduct tests and measurements at a wide variety of locations, thus providing realistic testing in your given scenario.

Machine Learning Expertise

Having run numerous successful machine learning projects, Fraunhofer IIS is well advanced in data recording, data organization and data storage, as well as in exploratory data analysis and the further processing of data using machine learning methods.

R&D Partners

With its in-depth experience in a wide range of technologies and artificial intelligence plus its expertise in measurement campaigns, Fraunhofer IIS is a reliable partner for joint German and international R&D projects.