Automated Vehicle Sensor Data Processing Using Map-Based Region-of-Interest
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Automated vehicle systems face high processing and communication costs due to the need to process and analyze vast amounts of data from sensors covering a 360° field of view, which is inefficient and costly.
Innovation Solution
A system that defines a region-of-interest (ROI) within the field of view based on roadway characteristics, using a digital map to focus processing resources and sensor data acquisition on specific areas, such as highways or intersections, while reducing processing and communication demands by adjusting sensor parameters and algorithms accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If sensor data from the entire 360° field of view is processed, then object detection coverage is improved, but processing power and communication costs increase significantly
Solution Approach 1:
The patent segments the 360° field of view into multiple regions based on roadway characteristics (e.g., high-risk areas like intersections and low-risk areas like straight highways). Processing resources are then allocated differently to each segment, with higher processing power dedicated to high-risk regions where objects are more likely to be detected, while reducing processing power for low-risk regions. This segmentation resolves the contradiction by maintaining comprehensive detection coverage through strategic focus on critical areas rather than uniform processing of all areas.
Solution Approach 2:
The patent applies local quality by adjusting processing parameters and algorithms according to the specific characteristics of each roadway region. High-risk regions receive enhanced processing quality with higher resolution and more sophisticated detection algorithms, while low-risk regions use simplified processing. This local differentiation maintains reliable object detection in critical areas while significantly reducing overall processing power consumption compared to uniform high-quality processing of the entire field of view.
2Reliability
If sensor data from the entire 360° field of view is processed, then object detection coverage is improved, but communication requirements and costs increase
Solution Approach 1:
The patent segments the field of view into regions and transmits only processed information from high-risk regions at high priority, while reducing or eliminating transmission of data from low-risk regions. This segmentation approach maintains effective communication for critical areas where objects are most likely to be detected, while significantly reducing the overall volume of data that must be communicated compared to transmitting all sensor data from the entire 360° field of view.
Solution Approach 2:
The patent extracts and prioritizes only the most critical information from sensor data based on roadway characteristics. By identifying and extracting high-risk regions where objects are most likely to be detected, the system transmits only this extracted information at high priority, while filtering out or reducing transmission of less critical data from low-risk areas. This extraction approach maintains effective object detection coverage through selective communication while reducing overall communication requirements and costs.
3Device complexity
If processing resources are allocated uniformly across all areas, then processing simplicity is maintained, but processing efficiency decreases
Solution Approach 1:
The patent implements dynamic processing resource allocation where the processing system automatically adjusts its behavior based on real-time roadway characteristic data. The system dynamically shifts processing focus between different regions of the field of view, concentrating resources on high-risk areas such as intersections while reducing resources for low-risk areas like straight highways. This dynamic adaptation maintains processing efficiency by allocating resources where they are most needed, while managing system complexity through automated decision-making based on pre-defined risk criteria.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An object-detection system (10) for an automated vehicle includes an object-detector (20), a digital-map (14), and a controller (12). The object-detector (20) is used to observe a field-of-view proximate to a host-vehicle. The digital-map (14) is used to indicate a roadway-characteristic (56) proximate to the host-vehicle. The controller (12) is configured to define a region-of-interest within the field-of-view based on the roadway-characteristic (56), and preferentially-process information from the object-detector (20) that corresponds to the region-of-interest.