Sensor Recognition Integration Stabilizing Object Coordinates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In autonomous driving systems, the integration of sensor data from various sensors leads to varying object coordinates, causing rapid changes and erroneous determination when sensor combinations change, which affects the accuracy of autonomous driving plans.
Innovation Solution
A sensor recognition integration device that includes a sensor object information unit for integrating data from multiple sensors, a prediction update unit for predicting object positions, an association unit for estimating associations, an offset parameter update unit for adjusting sensor data, and an integration update unit for generating stable integration object information, ensuring consistent object tracking across sensor changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If sensor information is integrated in accordance with the conventional technique, then the integration of multiple sensor data is achieved, but the coordinates of integrated object move rapidly when sensor combination changes
Solution Approach 1:
The system performs preliminary actions by predicting the future position of the object before integrating new sensor data. The prediction update unit calculates where the object will be at the next time point based on current motion state, and this predicted position is used as a reference to guide the integration process, preventing rapid coordinate changes when sensors are added or removed.
Solution Approach 2:
The patent introduces an intermediary mechanism - the prediction object - that acts as a mediator between different sensor observations. Instead of directly integrating raw sensor coordinates which may differ, the system uses the predicted object position as an intermediary reference point, and adjusts sensor coordinates relative to this prediction, thereby stabilizing the integrated object coordinates.
2Measurement precision
If multiple sensors are combined to detect objects, then detection coverage and accuracy are improved, but the complexity of examining specifications increases
Solution Approach 1:
The system changes parameters by transforming sensor object information into a standardized format with unified coordinate systems and time synchronization. The integration object information storage unit stores data with standardized parameters (predicted position, sensor coordinates, time stamps), allowing different sensor types to be integrated without examining each sensor's specific specifications individually.
3Speed
If sensor coordinates are used directly for autonomous driving determination, then real-time response is achieved, but erroneous determination occurs due to coordinate variations
Solution Approach 1:
The system implements feedback by continuously comparing actual sensor coordinates with predicted coordinates and using this information to adjust the integration process. The association unit uses the difference between predicted and actual positions to determine how to weight and combine sensor data, ensuring that the integrated result remains accurate and stable while maintaining real-time response capability.
Data Source
AI summary
The present invention provides a sensor recognition integration device capable of preventing a rapid change in coordinates of an integrated object and preventing, for example, an erroneous determination in an autonomous driving plan determination device, even when the combination of sensors that perform detection changes. In the present invention, since an object position is estimated in a state where information of a position detected by a sensor that recognizes an object in an external field is corrected or changed, the rapid change of coordinates of an integrated object is prevented even when the combination of sensors that perform detection changes.