Object Positioning Using Map-Sensor Heading Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous driving systems face errors in object positioning due to accumulated inaccuracies in localization, leading to potential collisions, lane deviations, and incorrect route settings, especially in low-precision map environments.
Innovation Solution
An object positioning method that combines map-based and sensor-based heading angles, using dead reckoning to determine the current position of an object by generating a second heading angle based on the comparison and compensation between map-based and sensor-based angles, with weights applied to improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If dead reckoning is used for position estimation using acceleration and direction sensors, then the system can operate without GPS, but accumulated errors in heading angle cause position drift and localization inaccuracy
Solution Approach 1:
The patent combines map-based heading angle (derived from road geometry and waypoints) with sensor-based heading angle (from direction sensors) to create a fused heading angle. This merging of independent heading angle sources compensates for the weaknesses of each individual method, preventing accumulated errors while maintaining reliability in GPS-denied environments.
Solution Approach 2:
The system continuously compares the map-based heading angle with the sensor-based heading angle and uses the difference to correct accumulated errors in real-time. This feedback mechanism detects drift and compensates for it by adjusting the position estimation based on the heading angle discrepancy, thereby maintaining positioning precision over time.
2Device complexity
If sensor-based heading angle alone is used for dead reckoning, then the system is simple to implement, but sensor inaccuracies accumulate over time causing large positioning errors
Solution Approach 1:
The map-based heading angle serves as an intermediary reference that mediates between the sensor-based heading angle and the true heading angle. By using road geometry and waypoint information as an intermediate source of directional information, the system corrects sensor drift without requiring direct measurement of true north or external references, thus maintaining simplicity while improving precision.
3Measurement precision
If map-based heading angle is used to correct sensor errors, then positioning accuracy improves, but the system requires access to precise map data and waypoints
Solution Approach 1:
The system performs preliminary actions by pre-processing map data to extract road geometry and generate waypoints before runtime. By preparing the map-based heading angle information in advance and storing it as structured data, the system reduces real-time computational complexity while maintaining high positioning precision during actual operation.
Data Source
AI summary
An object positioning method and apparatus is disclosed. The object positioning apparatus may obtain a reference position of an object, obtain a map-based heading angle of the object based on waypoints on a map, and determine a current position of the object based on the reference position and the map-based heading angle.