Radar Range-Doppler Detection of Small Stationary Objects
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional radar systems struggle to efficiently detect and classify smaller stationary objects due to sparse point cloud representations, leading to slower response times and decreased safety in autonomous driving applications.
Innovation Solution
Utilizing low-level radar data in the form of range-Doppler maps, range-azimuth maps, and other electromagnetic spectrum data, processed without thresholds, to feed an end-to-end deep convolutional detection and classification network for improved object detection and classification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional radar systems use point cloud representations for object detection, then the system structure is simple, but smaller stationary objects cannot be detected accurately and response time is slow
Solution Approach 1:
The patent applies preliminary action by performing preprocessing on raw radar signals to generate low-level spectrum data (range-Doppler maps, range-azimuth maps) before feeding into the deep convolutional network. This preprocessing step extracts and organizes relevant features in advance, enabling the network to process data more efficiently and achieve faster response times while maintaining high detection accuracy for small stationary objects.
Solution Approach 2:
The patent transitions from traditional point cloud representations to low-level spectrum data representations (range-Doppler maps and range-azimuth maps), effectively changing the dimensional structure of the input data. This dimensional transformation allows the deep convolutional network to exploit spatial and spectral features more effectively, improving both detection accuracy and response time for small stationary objects.
2Productivity
If conventional radar systems process data with thresholds and compressed representations, then computational load is reduced, but detection accuracy for small objects decreases
Solution Approach 1:
The patent changes the parameters of data representation by using low-level spectrum data (range-Doppler maps and range-azimuth maps) instead of compressed point cloud representations. This parameter change preserves more original signal information while maintaining computational efficiency, allowing the deep convolutional network to achieve high detection accuracy for small stationary objects without excessive computational burden.
3Measurement precision
If deep convolutional networks process raw electromagnetic signals directly, then detection accuracy improves, but computational complexity increases significantly
Solution Approach 1:
The patent applies preliminary action by preprocessing raw electromagnetic signals into low-level spectrum data (range-Doppler maps and range-azimuth maps) before inputting to the deep convolutional network. This preliminary processing organizes the data into structured formats that highlight relevant features, allowing the network to achieve high classification accuracy with reduced computational complexity compared to processing raw signals directly.
Solution Approach 2:
The patent transforms raw electromagnetic signals into two-dimensional spectrum representations (range-Doppler and range-azimuth maps), changing the data dimensionality. This transformation enables the deep convolutional network to efficiently process the data using spatial feature extraction, reducing computational complexity while maintaining or improving classification accuracy for stationary objects.
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
Enhances computational efficiency, allowing for quicker and more accurate detection and classification of stationary objects, thereby improving the performance and safety of autonomous driving systems.
Implementation Method 1
Raw electromagnetic signals reflected off stationary objects and received by a radar system
Implementation Method 2
The preprocessing may also filter non-stationary range-Doppler bins. The remaining low-level spectrum data represents stationary objects present in a field-of-view (FOV) of the radar system
Data Source
AI summary
This document describes techniques and systems for stationary object detection and classification based on low-level radar data. Raw electromagnetic signals reflected off stationary objects and received by a radar system may be preprocessed to produce low-level spectrum data in the form of range-Doppler maps that retain all or nearly all the data present in the raw electromagnetic signals. The preprocessing may also filter non-stationary range-Doppler bins. The remaining low-level spectrum data represents stationary objects present in a field-of-view (FOV) of the radar system. The low-level spectrum data representing stationary objects can be fed to an end-to-end deep convolutional detection and classification network that is trained to classify and provide object bounding boxes for the stationary objects. The outputted classifications and bounding boxes related to the stationary objects may be provided to other driving systems to improve their functionality resulting in a safer driving experience.