MaaS Demand Prediction With TEE for Privacy-Preserving Collaboration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
MaaS operators face challenges in collaborating with different transportation service providers due to fragmented IT systems, privacy concerns, and regulatory hurdles, leading to incomplete user insights and inefficient service optimization.
Innovation Solution
A privacy-preserving data-sharing framework using federated learning and Trusted Execution Environment (TEE) enables collaboration among MaaS operators, allowing secure access to fragmented context databases without merging sensitive data, facilitating dynamic partnerships and improved service delivery.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If TSPs share data to create integrated user experience, then user experience and service optimization are improved, but data privacy and security concerns worsen
Solution Approach 1:
A privacy-preserving data-sharing framework acts as an intermediary between TSPs, enabling data exchange while protecting sensitive information. The framework includes a data sharing platform that mediates communications, allowing TSPs to access fragmented context databases without directly sharing raw data, thus resolving the contradiction between data sharing benefits and privacy risks
Solution Approach 2:
Instead of sharing original sensitive data, the system creates and shares copies or representations of data (such as aggregated statistics, anonymized datasets, or data summaries) that preserve analytical value while removing personally identifiable information, enabling user experience improvement without compromising individual privacy
2Adaptability or versatility
If TSPs collaborate to expand service coverage, then service availability is improved, but IT system compatibility and integration complexity worsen
Solution Approach 1:
The IT systems are segmented into modular, standardized components that can be independently implemented by each TSP while maintaining compatibility through common interfaces. The data sharing framework provides standardized protocols and APIs that segment the integration complexity from individual TSP systems, allowing service expansion without requiring complete system harmonization
Solution Approach 2:
A universal data sharing platform and standardized communication protocols are implemented that can accommodate multiple TSP systems with different underlying technologies. This multi-functional framework enables TSPs to collaborate and expand service coverage while maintaining their existing IT architectures, reducing integration complexity through standardization
3Measurement precision
If data is merged to improve prediction accuracy, then service optimization is improved, but data privacy and regulatory compliance worsen
Solution Approach 1:
The system merges data in a controlled manner through the privacy-preserving framework, combining information from multiple TSPs to improve prediction accuracy while maintaining data separation and privacy protections. The merging process occurs at aggregated or anonymized levels, allowing improved measurement precision without violating regulatory requirements for data protection
Data Source
AI summary
Methods, systems, and computer programs are presented for sharing information among Mobility as a Service (MaaS) providers to improve service delivery while protecting data privacy. A first MaaS system receives, from a second MaaS system, a request to predict demand generated by the first MaaS system for user trips that transfer from the first MaaS system to the second MaaS system. The first MaaS system predicts, in response to the request, the demand generated by the first MaaS system using a predictor model executing at the first MaaS system on a trusted execution environment (TEE) to access data of the first MaaS system. The first MaaS system sends the predicted demand to the second MaaS system.