Neural Network Road Attribute Prediction System
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for updating digital maps, such as OpenStreetMap, rely heavily on manual human annotations, resulting in incomplete and inaccurate data due to high costs, leading to misleading routing decisions and inaccurate arrival time predictions for ride-hailing services.
Innovation Solution
A method using a data processing system with a neural network to predict road attributes by extracting trajectory features and map features from GPS data and image data, respectively, and classifying them into prediction probabilities to improve map data completeness and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual human annotations are used to update digital maps, then map data accuracy can be maintained through human expertise, but the cost increases significantly and data completeness decreases
Solution Approach 1:
The system enables map data to update itself automatically by using trajectory data from vehicles and satellites to infer road attributes without human intervention. The neural network processes raw trajectory data to automatically update map data, making the system self-sufficient and eliminating the need for costly manual annotations while improving data completeness.
Solution Approach 2:
The patent replaces the mechanical process of manual human annotation with an automated computational system using neural networks. The system substitutes human labor with machine learning algorithms that process trajectory data and satellite imagery to infer road attributes, thereby maintaining accuracy while significantly improving productivity and data completeness.
2Measurement precision
If manual annotations are used for map updating, then data accuracy can be ensured, but the time consumption and cost increase
Solution Approach 1:
The system performs preliminary actions by continuously collecting and processing trajectory data from vehicles and satellite data in real-time. This ongoing preliminary processing allows the system to have map data ready for immediate use without requiring time-consuming manual annotation processes when updates are needed, thus reducing updating time while maintaining accuracy through the trained neural network.
Solution Approach 2:
The patent replaces the time-consuming manual annotation process with automated neural network processing. The system uses machine learning algorithms to rapidly analyze trajectory data and satellite imagery, substituting slow human processes with fast computational methods that maintain accuracy while dramatically reducing the time required for map updates.
3Productivity
If automated methods are used to predict road attributes, then cost and time are reduced, but measurement precision may decrease compared to manual annotations
Solution Approach 1:
The system achieves universality by using a single neural network framework that processes multiple types of input data (trajectory data from various vehicles, satellite imagery, map data) to predict multiple road attributes simultaneously. This multi-functional approach allows the automated system to match or exceed manual annotation accuracy across different attribute types while maintaining high productivity and updating efficiency.
Solution Approach 2:
The patent combines multiple data sources and processing techniques into a composite predictive system. The neural network integrates trajectory data, satellite imagery, and existing map data to create a composite input that enhances prediction accuracy. This composite approach allows the automated system to achieve measurement precision comparable to manual annotations while maintaining the benefits of automation in terms of cost and time efficiency.
4Loss of information
If more road attributes are annotated manually, then map completeness improves, but the cost increases proportionally
Solution Approach 1:
The system enables self-service by automatically inferring multiple road attributes simultaneously from trajectory data and satellite imagery without requiring proportional manual annotation efforts. The neural network processes the same input data to predict multiple attributes (road type, speed limits, lane information, etc.), achieving improved map completeness while avoiding the proportional increase in annotation costs that would result from manual methods.
Data Source
AI summary
The disclosure relates to a method of predicting one or more road attributes. The method may include providing trajectory data of a geographical area. The method may further include providing map data, wherein the map data may include image data of the geographical area. The method may further include extracting trajectory features from the trajectory data and extracting map features from the map data. The method may further include using at least one processor to predict road attributes by inputting the trajectory features and the map features in a neural network and by classifying an output of the neural network into prediction probabilities of the road attributes. The disclosure also relates to a data processing system; to a non-transitory computer-readable medium storing computer executable code; and to a method of training an automated predictor.