Autonomous Vehicle Radar Mapping for Static Object Suppression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face performance issues due to false positive data from radar sensor systems caused by static metallic objects, which can lead to unnecessary stops or maneuvers, as these objects strongly reflect radar signals and create blind spots, impacting the vehicle's navigation and operation.
Innovation Solution
The use of a prior radar space map that includes data on predefined static objects allows the autonomous vehicle to generate a score representing the likelihood of a tracked object's presence, enabling the vehicle to suppress data from these objects and adjust its movement to reposition itself and restore occluded fields of view, thereby improving navigation and reducing false positives.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the radar sensor system detects static metallic objects, then the detection sensitivity is improved, but false positive data is generated causing unnecessary stops or maneuvers
Solution Approach 1:
The system performs preliminary classification of detected objects as static or dynamic before generating navigation commands. By预先 identifying that certain detected objects are static metallic structures rather than moving vehicles, the system avoids unnecessary stops while maintaining detection sensitivity for actual hazards
Solution Approach 2:
Instead of treating all detected objects as potential hazards requiring avoidance, the system inverts the approach by assuming detected objects are safe until proven otherwise through classification. Static objects are identified and excluded from hazard assessment, allowing the vehicle to maintain normal navigation without unnecessary maneuvers
2Reliability
If the autonomous vehicle stops or maneuvers around detected objects, then safety is improved, but productivity decreases due to unnecessary stops
Solution Approach 1:
The system performs preliminary classification to distinguish static metallic objects from moving vehicles before executing navigation maneuvers. This advance identification allows the vehicle to maintain continuous movement through areas with static objects, preserving productivity while maintaining safety through proper hazard assessment
Solution Approach 2:
The system extracts and separates the classification of static objects from the hazard response pathway. By removing static metallic objects from the set of objects requiring avoidance maneuvers, the system eliminates unnecessary stops and maintains navigation efficiency while still responding appropriately to actual hazards
3Device complexity
If the radar sensor system maintains fixed field of view, then device complexity is reduced, but blind spots are created by occluding static objects
Solution Approach 1:
The system dynamically adjusts the radar field of view based on the presence of occluding static objects. When a static object is detected that creates a blind spot, the radar actively steers to alternative angles to maintain detection coverage, balancing the simplicity of a fixed system with the information loss of blind spots
Solution Approach 2:
The system introduces an intermediary object detection and classification layer between the fixed radar and the navigation system. This intermediary identifies occluding static objects and mediates by adjusting radar viewing angles or compensating for blind spots through alternative detection methods, maintaining coverage without requiring complex active radar steering
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
This approach effectively suppresses data from static objects, enhancing the autonomous vehicle's navigation by reducing false positives and maintaining smooth operation, even in environments with static metallic objects, by utilizing prior data to adjust its path and sensor views.
Implementation Method 1
a radar sensor system, a lidar sensor system, and an image (camera) sensor system
Implementation Method 2
the static, metallic objects can strongly reflect radar signals
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
Various technologies described herein pertain to controlling an autonomous vehicle (100) to suppress data corresponding to predefined static objects (304, 606) in a radar output generated by a radar sensor system (102). A computing system (112) of the autonomous vehicle retrieves prior data for a geographic location from a prior radar space map (120). The prior radar space map includes prior data for geographic locations in an environment corresponding to whether predefined static objects to be suppressed in radar outputs are located at the geographic locations. The computing system generates a score representative of a likelihood of a tracked object being at the geographic location based on data from the radar output for the geographic location, data from an output of a second sensor system (104) for the geographic location, and the prior data for the geographic location from the prior radar space map. An engine (106), braking system (108), and/or steering system (110) are controlled based on the score.