ActiLedger: Privacy-Enhanced Multimodal Human Activity Recognition Using Blockchain in Crowdsensing

Abstract

The demand for secure and reliable data sharing has grown rapidly with the expansion of distributed systems and mobile sensing applications. This work presents a blockchain-based cloud framework that enables secure interaction between stakeholders and mobile devices. The system is designed with a simple web interface where users can register, log in, and access various features without difficulty. During registration, user information is stored in the blockchain along with transaction details such as block number, hash value, and timestamp, which helps maintain data integrity and traceability. After logging in, stakeholders can create sensing tasks using specific keywords. These tasks are recorded in the blockchain, allowing multiple mobile devices to participate and upload relevant sensed data. The uploaded data is stored securely in encrypted form within the cloud server. A keyword-based search mechanism is provided so that stakeholders can retrieve only the required data using valid inputs, while incorrect inputs return no results. Performance is analyzed using graphs that show block computation time and search efficiency with cache memory. Overall, the system offers a practical, secure, and scalable solution for decentralized data sharing.