Private decentralized crowdsensing with asynchronous blockchain access

Because of the popularity of mobile devices, crowdsensing has emerged as a data sensing paradigm for collecting large-scale data. Blockchain technology is promising to address problems of privacy and trust existing in centralized crowdsensing systems. However, it is challenging for crowdsensing applications to use a public blockchain to collect real-time or large-scale data since current blockchain systems lack of scalability. We propose a crowdsensing scheme that combines Trusted Execution Environments (TEE) with a public blockchain, achieving high efficiency with guarantees of privacy and trust. The scheme is a layer-two blockchain solution that supports off-chain multi-round sensing-data evaluations inside TEE enclaves, so that the sensing data needs not be propagated over the blockchain network. Besides, the scheme prevents false-reporting and free-riding for workers and requesters without reliance on a trusted third party. Moreover, the scheme secures the sensing data, letting only the worker and the corresponding requester know the data. Evaluations of our experimental prototype demonstrate the efficiency of our designs, as well as the reasonable on-chain monetary cost of running a task's smart contract and performing token payments using Ethereum.