Method for Extracting a Semantic Parameter

This invention is designed to search radar data for useful information. Specifically, it assists data analysis by producing a "semantic parameter," which is a quantitative value connected to a specific type of object. For example, it might take a data set recorded from a walking person and yield specific values for the speed of the torso or the leg. The goal of the invention is to make radar data easily legible according to previously specified terms.

The method calls for the determination of a velocity envelope for different objects. Whoever programs the detector must have some idea about the type of objects to be detected. For example, they must introduce some basic differentiation, such as the difference between upper and lower body. Different elements of the object are subsequently sorted and classified according to existing parameters. Data can then be accessed in terms of these parameters; leg speed, for example, can be analyzed as a separate variable. The invention can also be used to analyze relations between still objects.


  • Detection of moving speed of the human body
  • Analysis of individual moving parts
  • Determination of how a human body is moving: i.e., walking, running, bicycling, swimming, or other types of locomotion
  • Detect other types of movement: vehicles, animals, or any other object with specified parameters

Technical Summary

The invention calls for a radar signal which collects data about nearby moving objects. This data is then fed into a program, which classifies values according to specified semantic values. The method defines a range of values which it interprets as relevant. Within this range, the average speed is taken as the speed of the primary object (i.e., the torso, in the case of the human). Other objects are discerned on the basis of their different frequencies. I.e., swinging arms are identified because they move forward and backward quickly.

License - Stage of Development


License - Proposal

licensing opportunity and research collaboration

License - TNO reference number


We use anonymous cookies to enhance the use of our site.