Radar Filter Optimization Using Lidar Obstacle Points
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional laser scanning systems for autonomous mobile robots are costly due to high processing requirements and are susceptible to errors in adverse weather conditions, while radar systems offer lower resolution and are prone to noise, and combining radar and lidar data increases expense and noise.
Innovation Solution
Optimize radar filter parameters using lidar obstacle points to enhance accuracy and reduce false positives and negatives, employing signal-to-noise, spatial, and temporal filters, with optimization performed periodically in ideal conditions to improve radar data processing efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If laser scanning systems are used for obstacle detection, then measurement precision is improved, but device cost increases due to high processing requirements
Solution Approach 1:
The patent uses radar as an intermediary sensor to perform initial obstacle detection and filtering. Radar data serves as a preliminary filter before lidar data is processed, reducing the volume of lidar data that requires expensive high-precision processing while maintaining overall detection accuracy.
Solution Approach 2:
The patent segments the obstacle detection process into multiple stages: radar-based preliminary detection, spatial filtering, temporal filtering, and selective lidar verification. This segmentation allows low-cost radar to handle the bulk of detection while expensive lidar is used only when necessary.
2Measurement precision
If laser scanning systems are used for obstacle detection, then measurement precision is improved, but reliability deteriorates in adverse weather conditions
Solution Approach 1:
The patent merges radar and lidar sensor systems to combine their complementary strengths. Radar provides reliable detection in adverse weather conditions (rain, fog, smoke) while lidar provides high-precision detection in clear conditions. The system dynamically switches between or combines their outputs based on environmental conditions.
3Device complexity
If radar systems are used for obstacle detection, then device cost is reduced, but measurement precision deteriorates due to lower resolution and noise
Solution Approach 1:
The patent employs spatial filters and temporal filters as intermediary processing stages to clean up radar data. These filters act as mediators that remove noise and false detections from the low-cost radar system, improving its effective precision without requiring expensive hardware upgrades.
4Measurement precision
If radar and lidar data are combined for obstacle detection, then measurement precision is improved, but device cost increases and noise increases
Solution Approach 1:
The patent segments the data fusion process to avoid combining all radar and lidar data. Instead, it uses radar data only for spatial and temporal filtering operations, and combines it with lidar data only when lidar detects potential obstacles. This selective combination reduces noise and cost while maintaining precision.
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 radar obstacle detection accuracy in adverse weather conditions while reducing costs by optimizing radar filter parameters using lidar data, ensuring reliable obstacle detection for autonomous systems.
Implementation Method 1
radar sensor data is captured from a radar sensor on a moving machine
Implementation Method 2
Lidar sensor data is captured from a lidar sensor on the moving machine
Data Source
AI summary
Disclosed are techniques for automatically optimizing radar filter parameters. In embodiments, radar sensor data is captured from a radar sensor on a moving machine/vehicle. The radar sensor data is filtered using radar filter parameters to produce filtered radar sensor data. Radar obstacle points are produced from the filtered radar sensor data. Lidar sensor data is captured from a lidar sensor on the moving machine. Lidar obstacle points are produced from the lidar sensor data. The radar filter parameters are optimized using the lidar obstacle points.