Camera-Based Maps with Sensor Reflection Layers for Precise Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional systems that generate camera-based maps require significant computing resources due to the large amount of sensor data processing, and using only image data for localization can lead to reduced precision and accuracy in operations like localization, as they differ in feature extraction from other sensor data types such as RADAR and LiDAR.
Innovation Solution
Augmenting camera-based maps with sensor reflection information, such as RADAR or LiDAR data, by aligning and associating landmarks with secondary data to improve localization accuracy while reducing computational requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional systems generate map layers using RADAR data and LiDAR data, then measurement precision and localization accuracy are improved, but computing resources and processing requirements increase significantly
Solution Approach 1:
The patent segments the map into multiple layers, with a base layer generated from image data and separate overlay layers generated from RADAR and LiDAR data. This segmentation allows the system to process different sensor data types independently and selectively, reducing the computational burden of processing all sensor data simultaneously while maintaining localization accuracy through integrated multi-layer map representation.
Solution Approach 2:
The patent applies partial action by selectively generating map layers from secondary sensor data (RADAR, LiDAR) only in regions or contexts where enhanced localization precision is needed, rather than processing all sensor data universally. The system can choose to use only the base image layer or augment it with selective overlay layers based on operational requirements, reducing overall computing resource consumption.
2Measurement precision
If conventional systems provide multiple map layers to machines, then localization precision is improved, but network resources and data communication requirements increase
Solution Approach 1:
The patent extracts only the essential localization-relevant features from RADAR and LiDAR data to create condensed overlay layers, rather than transmitting complete raw sensor datasets. By extracting key geometric and reflective characteristics needed for localization while discarding redundant information, the system reduces data volume for communication while preserving localization precision.
Solution Approach 2:
The base image layer serves multiple functions: it provides the fundamental map structure for navigation and can be independently used for localization when secondary sensor data is unavailable. This multi-functionality reduces the need to continuously transmit all map layers, as the base layer alone can satisfy many operational requirements, thereby reducing network resource consumption.
3Device complexity
If systems use only image data for camera-based maps, then computing resources are reduced, but localization precision and feature extraction accuracy deteriorate
Solution Approach 1:
The patent merges the base image layer with selective overlay layers from secondary sensors to create an augmented map that combines the computational efficiency of image-based mapping with the localization precision of multi-sensor data. The merging is performed selectively, combining only the necessary elements from each data type, thereby maintaining reduced computing resource requirements while improving localization accuracy through integrated feature extraction from multiple sensor modalities.
4Reliability
If systems generate comprehensive multi-sensor maps, then reliability of navigation operations is improved, but processing time and productivity decrease
Solution Approach 1:
The patent performs preliminary action by pre-processing image data to create the base map layer, which can be generated and stored independently before navigation operations begin. This preliminary processing of the primary sensor data allows the system to quickly retrieve and use the base layer during navigation without waiting for time-consuming processing of secondary sensor data, thereby improving processing speed while maintaining navigation reliability through the availability of a pre-prepared foundational map.
Data Source
AI summary
In various examples, augmenting camera-based maps with sensor reflection information for autonomous and/or semi-autonomous systems and applications is described herein. Systems and methods described herein may augment map data representing a camera-based map using another type of data, such as RADAR data. For instance, image data generated using one or more machines navigating within an environment may be used to determine locations associated with landmarks (e.g., objects, features, etc.) located within an environment. The sensor data generated using the machine(s) may then be processed to determine whether the landmarks are associated with sensor reflections or whether the landmarks are not associated with sensor reflections. Additionally, the camera-based map may then be updated to include at least the locations associated with the landmarks and indications of whether the landmarks that are associated with sensor reflections.