Perception Sensor Prioritization for Real-Time Vehicle Object Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automotive perception sensor processing methods, such as those using RADAR and LIDAR systems, face high computational demands due to noisy and uncertain sensor data, particularly in prioritizing object detections for advanced driver assistance systems and autonomous driving, where existing methods often rely solely on proximity, failing to effectively focus processing resources on the most relevant objects.
Innovation Solution
A method and processing unit that assign priority values to sensor detections based on a pre-selected prioritization scheme in a coordinate system, weighting objects ahead of the vehicle more heavily, allowing for optimized processing of RADAR data and enhancing safety by prioritizing objects according to driving conditions and vehicle parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensor sensitivity is increased and more sensors are used, then measurement precision is improved, but computational demand increases excessively
Solution Approach 1:
The patent extracts and processes only the most relevant detections from the full sensor data set. By implementing a prioritization mechanism that identifies and processes only high-priority detections (those with potential safety relevance) while discarding or deferring low-priority ones, the system maintains measurement precision for critical objects while dramatically reducing computational demand on embedded processing units.
Solution Approach 2:
The patent applies different processing quality levels to different detections based on their priority. High-priority detections receive full processing attention with detailed analysis, while low-priority detections receive reduced or no processing. This local differentiation of processing quality allows the system to maintain high measurement precision where needed while reducing overall computational burden.
2Reliability
If all detections are processed equally, then object detection completeness is improved, but processing time increases excessively
Solution Approach 1:
The patent performs preliminary prioritization of detections before full processing. By pre-identifying which detections are most likely to be safety-relevant based on simple criteria (such as detection strength, object type, or location), the system can allocate processing time efficiently - ensuring complete and reliable processing of high-priority detections while using minimal or no processing resources on low-priority detections, thereby reducing overall processing time.
Solution Approach 2:
The patent applies partial processing action to detections based on priority. Rather than processing all detections to the same level of detail, the system applies full processing action only to high-priority detections that require complete analysis for safety reasons, while applying reduced or no processing action to low-priority detections. This selective partial action maintains detection reliability for critical objects while dramatically reducing processing time.
3Area of stationary object
If processing resources are allocated to distant objects, then detection coverage is improved, but processing of nearby objects may be insufficient
Solution Approach 1:
The patent applies different processing accuracy levels to detections based on their spatial priority and safety relevance. Nearby objects that pose immediate safety risks receive high-processing accuracy with detailed analysis, while distant objects receive reduced processing accuracy or are processed only for basic detection purposes. This local differentiation ensures that processing resources are concentrated on nearby objects where high accuracy is critical for safety, while still maintaining some level of detection coverage for distant objects.
Data Source
Figure 1~3
AI summary
Method for processing perception sensor data in an ego vehicle, comprising: receiving sensor data associated with locations of detections in a coordinate system having a y-axis and an x-axis. A priority value is assigned to the detections based on their location in the coordinate system. The priority values assigned to detections vary based on the based on the corresponding location of the detections in the coordinate system according to a pre-selected prioritisation scheme. Output data is then generated based on the detections and the assigned priority values.