Vehicle Object Relevance Field for Real-Time ADAS Prioritization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing advanced driver assistance systems (ADAS) face challenges in efficiently prioritizing and managing computational resources to evaluate objects in the environment of a vehicle, leading to suboptimal decision-making in real-time driving scenarios.
Innovation Solution
The use of a potential field generated by a computer system to determine relevance scores for objects based on their position, speed, and environmental layout, allowing for continuous evaluation and prioritization of objects without re-evaluation, thereby optimizing resource allocation and decision-making.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional object evaluation methods are used in ADAS, then all objects are evaluated equally, but computational resources are wasted on low-priority objects and insufficient on high-priority ones
Solution Approach 1:
The patent applies local quality by creating different evaluation criteria for different spatial regions around the vehicle. Objects in critical zones (front, rear, side zones) are assigned different relevance weights based on their position. This allows the system to focus computational resources on evaluating objects in high-risk areas with higher detail while using simpler evaluation for objects in low-risk areas, thereby improving decision-making accuracy without proportionally increasing overall computational complexity.
Solution Approach 2:
The patent changes the parameter of object evaluation from uniform to position-dependent relevance scoring. By introducing spatial parameters (distance, angle, zone classification) as variables in the evaluation function, the system dynamically adjusts the priority and depth of evaluation for each object based on its location relative to the vehicle. This parameter change enables efficient resource allocation where computational effort is proportional to the actual risk posed by each object.
2Reliability
If comprehensive object evaluation is performed for all objects, then decision accuracy is maintained, but real-time processing becomes difficult due to high computational overhead
Solution Approach 1:
The patent segments the environment into multiple zones (front zone, rear zone, side zones, etc.) with different evaluation priorities. By dividing the computational task into zone-based segments, the system can process objects in parallel according to their zone's priority level. High-priority zones receive more detailed evaluation while low-priority zones use simplified assessment, enabling real-time processing while maintaining accuracy for critical objects.
Solution Approach 2:
The patent applies partial action by performing full comprehensive evaluation only on objects in high-priority zones, while using simplified evaluation for objects in low-priority zones. This selective approach ensures that decision-critical objects receive thorough assessment while reducing overall computational overhead, thereby achieving real-time processing speeds without sacrificing accuracy for the most important objects.
3Device complexity
If object priority is determined by distance only, then computation is simplified, but objects at same distance with different risks are treated equally
Solution Approach 1:
The patent introduces asymmetry in the evaluation system by treating objects in different angular positions and zones differently, even at the same distance. Rather than using symmetric distance-based evaluation, the system assigns different relevance weights to objects based on their specific zone (front, rear, left side, right side) and angular position. This asymmetric approach allows the system to recognize that an object at 30 degrees in the front zone poses different risk than an object at the same distance in the side zone, thereby improving assessment precision without significantly increasing algorithmic complexity.
Data Source
AI summary
A computer includes a processor and a memory, and the memory stores instructions executable by the processor to generate a potential field covering an environment surrounding an ego vehicle and centered on the ego vehicle, and determine a relevance score for an object in the environment according to a position of the object in the potential field. The potential field indicates relevance to the ego vehicle.