Transportation Data Lake for AI Training
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in creating robust training datasets for AI algorithms in transportation environments, as they often lack diversity, are outdated quickly, and require costly manual curation, while also facing connectivity and privacy issues.
Innovation Solution
A transportation environment data lake is established, leveraging existing road infrastructure solutions with multimodal sensors, compute platforms, and connectivity, to collect, process, and provide diverse, up-to-date data for AI training, while ensuring data privacy and security.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual data curation is used to create training datasets, then data quality and diversity can be improved, but costs and time consumption increase significantly
Solution Approach 1:
The system enables automated data collection, processing, and curation through AI algorithms that operate autonomously. Roadside infrastructure solutions automatically capture sensor data, edge services process and annotate it, and the platform delivers curated datasets without human intervention, making the system self-sufficient and eliminating manual curation costs
Solution Approach 2:
The patent introduces an intermediary automated processing layer between raw sensor data and training datasets. Edge services act as intermediaries that automatically collect, annotate, and process data from multiple sources, transforming raw data into curated training datasets without requiring manual human effort
2Productivity
If existing datasets are used for AI training, then immediate training can begin, but the data becomes outdated quickly and lacks diversity
Solution Approach 1:
The system performs preliminary data collection and processing actions continuously in advance. Roadside infrastructure solutions continuously capture sensor data and store it in data lakes, so when training is needed, fresh and diverse data is already prepared and immediately available, eliminating the outdated data problem while maintaining fast training speeds
Solution Approach 2:
The patent establishes continuous data collection, processing, and updating operations. The system continuously gathers sensor data from roadside infrastructure, continuously processes and annotates it through edge services, and continuously updates the data lake, ensuring the training data remains fresh and diverse without interruption
3Reliability
If sensor data is collected from roadside infrastructure, then data diversity and real-time accuracy are improved, but data privacy and security challenges arise
Solution Approach 1:
The system applies different processing qualities to different parts of the data. Sensitive personal information is anonymized and processed with higher security measures, while non-sensitive sensor data maintains its original quality for training purposes. This localized quality approach preserves data accuracy where needed while protecting privacy where required
Data Source
AI summary
Disclosed are embodiments that provide a transportation environment data service. The transportation environmental data service includes harvesting services that crawl roadside infrastructure solutions to obtain sensor data collected from sensors physically positioned at the roadside infrastructure. In some cases, the roadside infrastructure solutions perform additional processing on the sensor data. For example, some roadside infrastructure performs object detection and/or object recognition. When encountering these solutions, the edge or harvesting service also collects the object detection and/or object recognition information. Customers can subscribe to various data services provided by the transportation environment data service. For example, some subscribers indicate an interest in any updates of environmental data for a particular region. Other subscribers are interested in video data associated with any vehicular accidents detected by the transportation environment data service.