Radar Object Classification Using Dimensioning Information
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Radar sensors face challenges in accurately locating and classifying objects in their monitored area due to ambiguities in two- or multidimensional frequency representations, which can lead to incorrect classifications and inaccuracies in semi-automated driving systems, especially in distinguishing between objects of different sizes and distances.
Innovation Solution
The method involves converting radar signals into a two- or multidimensional frequency representation and supplying dimensioning information, such as distance and azimuth angle, to an artificial neural network (ANN) to improve object classification and location accuracy. This information can be derived from the frequency representation itself or from additional sensors like LIDAR, and is used to refine the classification process by reducing ambiguities and suppressing interfering signal components.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If radar signals are converted into frequency representation for object detection, then the distance and speed of objects can be determined, but ambiguities arise in classifying objects of different sizes and distances
Solution Approach 1:
The patent introduces a third dimension (spatial dimensioning information) to the existing two-dimensional frequency representation (distance-speed or distance-azimuth). By adding dimensioning information about the physical size and spatial extent of objects, the system resolves ambiguities between near-small and distant-large objects that appear similar in the original frequency representation.
Solution Approach 2:
The patent uses dimensioning information as an intermediary element that mediates between the ambiguous frequency representation and the final classification. This additional spatial information acts as a bridge that helps the neural network disambiguate between different object types by providing context about their physical dimensions.
2Measurement precision
If dimensioning information is added to resolve ambiguities, then classification accuracy improves, but system complexity increases
Solution Approach 1:
The patent makes the existing neural network architecture multi-functional by enabling it to process both the original frequency representation data and the additional dimensioning information. The same neural network structure that performs classification also integrates spatial dimensioning, eliminating the need for separate processing systems.
Solution Approach 2:
The dimensioning information is calculated and prepared in advance before being fed into the neural network for classification. By pre-processing the spatial dimensioning data from the frequency representation, the system reduces the computational complexity during the actual classification phase.
3Adaptability or versatility
If radar waves are used for object detection, then detection is independent of light conditions, but ambiguities occur due to radar waves reflecting on undesirable locations
Solution Approach 1:
The patent applies local quality analysis by examining the spatial distribution and dimensioning characteristics of reflected radar signals from different locations. By analyzing the local spatial properties and dimensioning information of each detected target, the system can distinguish between genuine objects and reflections from undesirable locations.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The integration of dimensioning information significantly enhances the accuracy of object classification and location, reducing incorrect classifications and improving the reliability of radar-based object detection in various environments, thereby supporting safer semi-automated driving.
Implementation Method 1
at least one radar sensor (1) having at least one transmitter (11) and at least one receiver (12) for radar waves
Implementation Method 2
the physical nature of the radar measurement, which as a measured variable ultimately detects which portion of the radiation that is emitted into the monitored area is reflected
Data Source
AI summary
A method is described for locating and/or classifying at least one object, a radar sensor that is used including at least one transmitter and at least one receiver for radar waves. The method includes: the signal recorded by the receiver is converted into a two- or multidimensional frequency representation; at least a portion of the two- or multidimensional frequency representation is supplied as an input to an artificial neural network, ANN that includes a sequence of layers with neurons, at least one layer of the ANN being additionally supplied with a piece of dimensioning information which characterizes the size and/or absolute position of objects detected in the portion of the two- or multidimensional frequency representation; the locating and/or the classification of the object is taken from the ANN as an output.