Autonomous Vehicle Image Transfer Using Map Difference Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in efficiently communicating and storing large volumes of image data, which is difficult and costly due to the high volume of data required for real-time communication and storage.
Innovation Solution
The system identifies relevant objects in images and transfers a list of these objects, along with their descriptions, allowing a remote server to synthesize an approximation of the image by rendering objects at their proper locations, and uses difference images derived from predefined maps to reduce data volume by removing redundant background information, enabling compression and reconstruction of the original image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If full image data is transferred from autonomous vehicle to remote server, then image representation accuracy is maintained, but data transfer volume and storage requirements become prohibitively large
Solution Approach 1:
The patent extracts only the essential semantic information from full images - specifically object identities, bounding boxes, and attributes - while discarding redundant pixel data. This extraction principle reduces data volume dramatically while preserving the critical information needed for remote monitoring and analysis.
Solution Approach 2:
Instead of transferring actual image pixels, the system creates simplified semantic copies representing the essential content - object detections with their properties. These semantic copies serve as efficient representations that maintain information quality while minimizing data transfer requirements.
2Measurement precision
If high-resolution images are captured and transferred in real-time, then detection precision is improved, but communication bandwidth requirements and latency increase
Solution Approach 1:
The system extracts only the critical detection results - object classes, positions, and key attributes - from the captured images. This extraction eliminates the need to transmit the full high-resolution image data, thereby reducing transfer time and latency while maintaining detection precision through the preserved semantic information.
Solution Approach 2:
The patent transforms the data representation from continuous pixel values to discrete semantic parameters (object IDs, bounding box coordinates, confidence scores). This parameter transformation reduces data complexity and transfer requirements while preserving the essential detection precision information.
3Loss of information
If complete sensor data is stored for later analysis, then data completeness is maintained, but storage costs and processing requirements become unsustainable
Solution Approach 1:
The system extracts and stores only the essential semantic content from sensor data - object detections, their properties, and spatial relationships. This extraction approach maintains the completeness of critical information while reducing storage requirements by eliminating redundant pixel and sensor raw data.
Solution Approach 2:
The patent creates compact semantic copies of sensor data that preserve the essential information needed for analysis. These semantic representations serve as efficient storage alternatives to full-resolution images and complete sensor datasets, maintaining analytical completeness while dramatically reducing storage volume.
Data Source
AI summary
A system includes at least one imaging sensor and a processor. The processor is configured to acquire, using the imaging sensor, detected data describing an environment of an autonomous vehicle. The processor is further configured to derive reference data, which describe the environment, from a predefined map, to compute difference data representing a difference between the detected data and the reference data, and to transfer the difference data. Other embodiments are also described.