In crowdsourcing, task allocation service helps to screen out suitable workers to realize more extensive perceptual data collection. However, due to the vulnerability of existing technologies, it is also subject to security and privacy issues. Although there exist some related works that enable to protect the privacy of both requesters and workers, they cannot well support flexible and expressive matching polices for multi-keyword in multi-worker scenarios.

Since the inner-product encryption (IPE) mechanism enables to support multiple types of constraint expressions, a construction of privacy-aware task allocation system based on two forms of IPE is proposed in this paper. In such a scheme, only the crowdsourcing service provider (CSP) has the ability to perform fine-grained inner-product policy matching over the encrypted position and other attributes. Moreover, the CSP could calculate the corresponding conversion ciphertext for workers who meet the conditions, thus reducing the amount of user calculation. And meanwhile, through the security proof, we prove the indistinguishability for task content and location information, respectively. Furthermore, the implementation analysis and evaluation show that the scheme is feasible and effective.

在众包中，任务分配服务有助于筛选出合适的工作人员，以实现更广泛的感知数据收集。但是，由于现有技术的脆弱性，它还受到安全和隐私问题的困扰。尽管存在一些相关的工作可以保护请求者和工作人员的隐私，但是它们不能很好地支持在多工作人员场景中针对多关键字的灵活且富有表现力的匹配策略。

由于内部产品加密（IPE）机制可以支持多种约束表达式，因此本文提出了一种基于两种IPE形式的隐私感知任务分配系统。在这种方案中，仅众包服务提供商（CSP）能够对加密位置和其他属性执行细粒度的内部产品策略匹配。而且，CSP可以为满足条件的工作人员计算相应的转换密文，从而减少了用户计算量。同时，通过安全性证明，我们分别证明了任务内容和位置信息的不可区分性。进一步的实施分析和评价表明该方案是可行和有效的。