Stereo-Assist Object Localization With Sparse Map Navigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in navigating effectively due to the vast amounts of data they need to process and store, particularly with traditional mapping technologies, which can limit their ability to accurately identify objects and navigate through environments.
Innovation Solution
A navigation system that utilizes multiple cameras to capture images from different angles, processes these images using trained models to determine object signatures, and combines the data to accurately locate objects within the vehicle's environment, employing a sparse map for efficient navigation without the need for extensive data storage or transfer.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional mapping technology is used for autonomous navigation, then comprehensive environmental data is available, but data storage requirements and processing complexity increase significantly
Solution Approach 1:
The patent extracts only the essential navigational features from the environment (lane markings, road signs, intersections, obstacles) and stores them as sparse map data points rather than storing complete environmental maps. This selective extraction reduces data storage requirements while maintaining navigation accuracy by focusing on critical path-related information.
Solution Approach 2:
The navigation system segments the environment into discrete navigational elements (lane markings, intersections, obstacles, road signs) and processes each segment independently. This segmentation allows the system to handle complex navigation scenarios by breaking down the overall navigation task into manageable components, reducing processing complexity.
2Measurement precision
If multiple cameras capture images from different angles for object detection, then object location accuracy improves, but data processing time and computational load increase
Solution Approach 1:
The system performs preliminary processing of images from multiple cameras by extracting key features and creating initial object detections before combining data. This preliminary action prepares the data in advance, making the subsequent fusion process more efficient and reducing overall processing time while maintaining accurate object location determination.
Solution Approach 2:
The patent merges object representations from multiple camera views by combining image data, depth information, and feature extractions into a unified object model. This merging process integrates information from different angles and sensors to improve object location accuracy while using efficient algorithms to minimize computational load and processing time.
Data Source
AI summary
Systems and methods for navigating a host vehicle are disclosed. In one implementation, a system includes a processor configured to receive a first image acquired by a first camera and a second image acquired by a second camera onboard the host vehicle; identify a first representation of an object in the first image and a second representation of the object in the second image; input to a first trained model at least a portion of the first image; input to a second trained model at least a portion of the second image; receive the first signature encoding determined by the first trained model and the second signature encoding determined by the second trained model; input to a third trained model the first signature encoding and the second signature encoding; and receive an indicator of a location of the object determined by the third trained model.