Multi-Technique Target Fusion for Autonomous Driving Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous driving systems using machine learning techniques may experience recognition errors due to inconsistencies in sensor information and stored environment features.
Innovation Solution
A target calculation method executed by a computing device that acquires sensor outputs, detects targets using multiple techniques (rule-based and AI-based), determines if targets are the same across different detection methods, and merges the target states to output a single, accurate target state.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single detection technique is used, then the detection process is simple and fast, but recognition errors occur due to inconsistencies in sensor information
Solution Approach 1:
The patent combines multiple detection techniques (rule-based detection and machine learning-based detection) to detect the same target. The rule-based detection unit applies predefined rules to sensor information, while the machine learning-based detection unit uses trained models. Both detection results are merged through a determination unit that identifies consistent targets across different techniques, thereby improving detection accuracy and reducing recognition errors while managing complexity through structured integration.
2Measurement precision
If multiple detection techniques are used, then recognition errors are reduced, but the processing time and computational load increase
Solution Approach 1:
The patent performs preliminary actions by pre-training machine learning models offline and preparing rule-based detection criteria beforehand. During actual target detection, the system directly applies these pre-prepared models and rules to sensor information, avoiding the need for real-time model training or complex rule generation. This preliminary preparation significantly reduces real-time processing time while maintaining high detection precision through multiple techniques.
3Reliability
If multiple detection techniques are used, then target detection accuracy is improved, but the system complexity increases
Solution Approach 1:
The patent segments the detection system into distinct functional units: a rule-based detection unit that applies predefined rules, a machine learning-based detection unit that uses trained models, and a determination unit that merges results. Each unit has a specific, simplified function, making the overall complex system manageable through modular design. This segmentation allows independent optimization and maintenance of each detection technique while achieving improved target detection reliability through their coordinated operation.
Data Source
AI summary
A target calculation method is a target calculation method executed by a computing device including an acquisition unit that acquires a sensor output that is an output of a sensor that acquires information on a surrounding environment, target calculation method including: detection processing of detecting a target by a plurality of techniques using the sensor output and detecting a target state including at least a position and a type of the target; same target determination processing of determining a same target from a plurality of the targets detected by each of the plurality of techniques in the detection processing; and merging processing of merging the target states of the target determined to be the same target in the same target determination processing and outputting the merged target states as a merged target.