Radar Point Clouds With Appended Features for Object Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Radar sensor systems conventionally discard energy values in bins other than those identified as corresponding to objects, limiting the information available for object detection and tracking, and do not transmit data tensors due to bandwidth constraints.

Innovation Solution

A radar system generates point clouds by identifying peak energy values, computing feature values from neighboring bins, and appending these features to the point cloud entries, which are then transmitted for object detection and tracking.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If the radar system transmits data tensors to the processing device, then the information available for object detection and tracking is improved, but the bandwidth requirements and transmission complexity increase significantly

Engineering Contradiction:

Improveinformation available for object detectionVSAvoidtransmission complexity

Core Design Contradiction:

Loss of informationVSDevice complexity

Solution Approach 1:

The patent extracts only the necessary feature information from the radar tensor data and appends it to the point cloud, rather than transmitting the entire data tensor. This extraction approach maintains object detection information while significantly reducing transmission bandwidth requirements and system complexity.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent transforms the data representation by appending feature values as an additional dimension to each point cloud entry. This allows rich information from the radar tensor to be conveyed through a compact point cloud structure, improving information availability without proportionally increasing transmission complexity.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

2Device complexity

If the radar system uses only peak energy bins for point cloud generation, then the processing simplicity is maintained, but the object detection accuracy deteriorates due to loss of useful information

Engineering Contradiction:

Improveprocessing simplicityVSAvoidobject detection accuracy

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The patent applies local quality by selectively processing different bins in the radar tensor. Instead of uniformly processing all bins or none, it identifies peak energy bins and selectively extracts features from their neighboring bins, applying different processing quality to different regions of the data based on their relevance to object detection.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent performs preliminary peak analysis to identify significant bins before generating the point cloud. This preliminary identification allows the system to focus computational resources on extracting features from relevant neighboring bins, improving detection accuracy while maintaining overall processing efficiency through selective rather than exhaustive processing.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS12571908B2Generating point clouds with appended features for use in perception

Publication Date: 2026.03.10 GM CRUISE HOLDINGS LLC

AI summary

The technologies described herein relate to a radar system that is configured to generate point clouds based upon radar tensors generated by the radar system. More specifically, the radar system is configured to identify bins in radar tensors that correspond to objects in an environment of the radar system, and to use energy values in other bins to construct a point cloud. A computing system detects objects in an environment of the radar system based upon the point clouds.