RTAF: A Reliable Task Allocation Framework for Enhancing Privacy, Data Quality, and Cost Efficiency in Mobile Crowdsensing

Abstract

Mobile crowdsensing is an emerging sensing paradigm that utilizes widely distributed mobile smart devices to rapidly collect data to support multiple applications in the Internet of Things. In mobile crowdsensing, collecting high-quality data at a low cost while protecting participants’ privacy is essential for developing high-quality applications and providing superior services. However, the paradigm faces many challenges, including protecting participants’ private information, improving data quality, and efficiently assigning tasks. To address these challenges, a Reliable Task Allocation Framework (RTAF) is designed to protect participant privacy, enhance data quality, and reduce costs. Specifically, the RTAF consists of three key algorithms: (1) A Bilateral Zero-bias Location Privacy-preserving (BZLP) algorithm is designed to protect the location privacy of tasks, ensuring the confidentiality and security of participants’ privacy. (2) An Unmanned Aerial Vehicle (UAV) Truth-driven Workers’ Trust degree Recognition (UAV-TWTR) algorithm is proposed to evaluate the data truth of tasks and identify trusted workers, reducing the risk of malicious workers attacking the system. (3) A Bi-objective Optimization Allocation of Tasks (BOAT) algorithm is developed to construct an optimal allocation scheme to ensure that the tasks can be assigned to appropriate workers, improving data collection accuracy and reducing the cost. Experimental results demonstrate that the RTAF surpasses the state-of-the-art method, enhancing the F1-score by 24.5%, increasing accuracy by 24.3%, and reducing the movement costs of workers by 19.1%.