Abstract
Mobile Crowdsourcing (MCS) leverages the sensing capabilities of smartphones but faces significant challenges regarding worker privacy and data trustworthiness. This paper proposes a novel security framework that ensures worker anonymity and location privacy without compromising data accuracy. By integrating certificate-less traceable ring signatures (CL-TRS) with a "current location privacy" model based on timed-release encryption (TRE), the scheme allows for high-precision data collection while protecting workers from location-based attacks. The design is optimized for mobile devices by eliminating resource-intensive bilinear pairings on the worker's side, making it both secure and computationally efficient.
1. Introduction
Mobile Crowdsourcing (MCS) has emerged as a powerful paradigm for distributed sensing, enabling applications like traffic monitoring and environmental sensing. However, widespread adoption is hindered by privacy concerns. Workers are often reluctant to share location-tagged data that could reveal sensitive information about their habits or home addresses. Conversely, requesters require assurance that the data is authentic and comes from authorized participants.
Traditional solutions like spatial cloaking often degrade data quality to protect privacy. This paper addresses these issues by proposing a scheme that maintains 100% data accuracy while ensuring worker safety through temporal privacy and lightweight cryptography. The authors set out to find a way to keep data accurate while keeping workers safe and their devices responsive.
2. System Architecture
The proposed MCS framework involves four primary entities: Workers, the MCS Server, Requesters, and a Time Server.
Workers collect data and submit signed, encrypted answers. The MCS server acts as an intermediary, while Requesters initiate tasks and receive results. A dedicated Time Server is introduced to facilitate timed-release encryption, ensuring that sensitive location data is only accessible after a predefined period.
3. Methodology
The core of the scheme relies on several cryptographic innovations designed to balance security and performance. The notations used in the formal construction of these protocols are summarized in the table below.
3.1 Certificate-less Traceable Ring Signature (CL-TRS)
This component provides anonymity by allowing a worker to sign an answer on behalf of a group (a "ring"). To a verifier, the answer appears to originate from any authorized member. Crucially, the scheme includes a traceability feature that can be triggered by the requester in case of malicious data submission, ensuring accountability without compromising privacy during normal operations. By using certificate-less cryptography, the authors avoid the "key escrow" problem where a single authority holds all master keys.
3.2 Timed-Release Encryption (TRE) and Current Location Privacy
The authors introduce the concept of "current location privacy," based on the assumption that location data is most sensitive while the worker is physically present at that site. By using TRE to "time-lock" the ciphertext ($CT$), the system ensures that precise location information is only revealed after a specified decryption time ($T$), once the worker has moved on. This preserves privacy without sacrificing the resolution of the collected answers ($m$).
3.3 Pairing-Free Operations
To accommodate the limited battery and processing power of smartphones, the scheme offloads heavy bilinear pairing operations to the server or initialization phase. The worker's device only performs simpler scalar multiplications, significantly reducing the computational burden compared to previous models.
4. Performance Evaluation
The scheme was evaluated using the JPBC (Java Pairing-Based Cryptography) library on a simulated environment. The authors tested the system with varying ring sizes (up to 30 members) to measure computational load.
- Efficiency: Encryption and decryption times for workers are independent of group size, typically ranging between 20ms and 24ms.
- Scalability: While signing and verification times scale linearly with the ring size, a ring of 25 members (providing robust anonymity) still requires less than 1.2 seconds for signing.
- Resource Optimization: By eliminating bilinear pairings on the worker side, the scheme significantly reduces the computational burden compared to previous traceable ring signature (TRS) and timed-release encryption (TRE) models.
5. Conclusion
The proposed scheme effectively balances the conflicting requirements of data precision and worker privacy in MCS. By leveraging the temporal dimension of privacy and optimizing for mobile hardware, it provides a practical solution for secure crowdsourcing. Future work will explore decentralizing the Time Server using blockchain technology and addressing privacy in task allocation, as well as testing the system's resilience against potential collusion between the MCS server and other entities.
Summary updated