TSWCrowd: A Decentralized Task-Select-Worker Framework on Blockchain for Spatial Crowdsourcing

Spatial crowdsourcing is an effective and novel method. In crowdsourcing systems, a centralized platform is traditionally used to allocate tasks and select workers. Centralized platforms always face following challenges: 1) How to ensure the rationality of tasks allocating; 2) How to ensure the payments of workers in the system when dishonest requesters exist; 3) How to ensure the maximum number of tasks are assigned. 4) How to ensure the integrity and reliability of the centralized platform. To solve these problems, this article proposed a distributed blockchain-based crowdsourcing framework - TSWCrowd (Task Select Worker Crowd). In this framework, tasks are sorted according to specific rules, thus tasks with higher priority are assigned to workers earlier. Workers who are available for a task will be selected and return a result. Then the deployed smart contracts will pay the basic payment automatically. At the same time, relevant contracts also calculate and pay the quality payment according to the proposed quality reward formulation. The proposed TSWCrowd framework on-chain involves a public dataset and uses solidity to compile the smart contracts. The framework was deployed on a local private blockchain. The decentralization property of the blockchain ensures the reliable assignment of tasks. Task-select-worker (TSW) algorithm sorts tasks to ensure reliability. In this paper, the proposed framework was compared with the ABCrowd auction mechanism on-chain and the VCG mechanism off-chain. The results show that the average distance is shorter and the payment is higher, thus reaches the reasonability, reliability and availability.