Radar Sequential Clustering for Low-Burden Object Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicle driving-control systems face a computational burden in processing vast amounts of radar sensor data to identify and track objects, leading to inefficiencies and potential errors due to limited computational resources.
Innovation Solution
The implementation of a method that clusters radar data points in two-dimensional spaces (range-Doppler and range-azimuth) using algorithms like DBSCAN, reducing the number of data points to be analyzed and improving tracking accuracy by forming overlap clusters that represent multiple data points, thereby simplifying the computational load and enhancing object detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If radar data points are processed individually to ensure accurate object detection, then measurement precision is improved, but computational burden increases significantly
Solution Approach 1:
The patent segments the computational task by dividing radar data processing into two stages: (1) clustering data points into groups representing potential objects, and (2) individually analyzing only these clustered groups. This segmentation reduces the number of individual data points requiring detailed analysis while preserving detection accuracy through the clustering process.
Solution Approach 2:
The patent merges multiple radar data points that represent the same physical object into single clusters. By combining these points based on spatial proximity and temporal consistency, the system reduces computational load while maintaining accurate object representation through the merged cluster entities.
2Measurement precision
If all radar data points are analyzed in detail to improve tracking accuracy, then object tracking precision is improved, but loss of time increases due to extensive processing
Solution Approach 1:
The patent performs preliminary clustering of radar data points before detailed tracking analysis. By pre-grouping data points into clusters based on their spatial and temporal characteristics, the system prepares the data structure in advance, allowing subsequent tracking operations to focus on clustered groups rather than individual points, thus reducing processing time while maintaining accuracy.
3Measurement precision
If the number of sensors is increased to improve environmental perception, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent creates a universal clustering framework that can process data from multiple sensor types (radar, LiDAR, cameras) using the same algorithmic approach. This multi-functional clustering system handles diverse sensor inputs through a unified processing pipeline, reducing overall system complexity despite the presence of multiple sensors by applying a single versatile processing methodology.
Data Source
AI summary
In one embodiment, a method includes accessing a set of data points captured using a radar system of the vehicle. Each data point is associated with at least three measurements include a Doppler measurement, a range measurement, and an azimuth measurement in reference to the radar system. The method also includes clustering the set of data points into one or more first clusters based on a first pair of the three measurements associated with each of the data points; and clustering the set of data points into one or more second clusters based on a second pair of the three measurements associated with each of the data points. The second pair being different from the first pair of the three measurements.