Object Recognition Fusion Data Discontinuity Handling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object recognition devices in autonomous vehicles face issues with unnatural object movement recognition due to excessive changes in fusion data from previous to current values, leading to inaccurate object recognition.
Innovation Solution
An object recognition device and method that includes a data reception unit, prediction processing unit, and tracking processing unit to generate prediction data and execute specialized tracking processing when data changes exceed a tolerable range, using a specific movement model to adjust physical quantities and prevent fusion data discontinuity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the number of sensors associated with fusion data changes, then the object recognition system can adapt to different detection scenarios, but the fusion data changes excessively leading to unnatural object movement
Solution Approach 1:
The prediction processing unit generates prediction data in advance by predicting the current value of fusion data using a previous value and a specific movement model. This preliminary prediction is then used as a reference when sensor configurations change, allowing the system to maintain stable fusion data by comparing actual sensor data against the predicted values and making appropriate adjustments.
2Productivity
If standard tracking processing is used, then the system processes fusion data efficiently, but excessive changes in fusion data cause inaccurate object recognition
Solution Approach 1:
The tracking processing unit implements feedback by comparing the actual fusion data with the prediction data generated by the prediction processing unit. When the change in fusion data exceeds a tolerable range, the system uses the prediction data as a reference to adjust the fusion data, ensuring that object recognition accuracy is maintained while still processing data efficiently through the tracking pipeline.
Data Source
AI summary
Provided is an object recognition device capable of preventing a change of fusion data from a previous value to a current value that exceeds a tolerable range. The object recognition device is configured to execute specialized tracking processing when a change of the fusion data from the previous value to the current value exceeds a tolerable range, and generate tracking data that is equivalent to prediction data in the specialized tracking processing by setting an adjustment physical quantity for solving a fusion data discontinuity state.