针对现有群智感知激励方法难以满足分布式环境下用户的隐私需求，该文提出一种基于Tangle网络的群智感知隐私保护激励方法TNIP（tangle network incentive policy）。首先，基于隐私保护思想，感知网络采用Tangle网络框架，分布式记录方式保障网络安全性；其次，为防止恶意竞争和区块链带来的PoW（proof of work）门槛等问题，设计了PoW控制算法，降低参与门槛和恶意竞争；接着，借助基于ECDSA（elliptic curve digital signature algorithm）的数字签名方法增强了感知数据保护，使数据更安全；然后，利用量化评估模型对隐私泄露风险进行量化评估，再通过可信的里程碑交易控制网络风险，增强感知网络隐私性；最后，利用真实数据集，通过仿真试验对该文提出的TNIP方法的激励效果和隐私保护方法进行对比分析，试验结果表明，TNIP方法在三类任务的平均参与率较CSII（cross-space multi-interaction-based-dynamic incentive scheme）方法和TCS（tangle-net crowd sensing）方法分别提高了8.79%和5.93%。

In response to the inadequacy of existing crowd sensing incentive methods to meet the privacy requirements of users in distributed environments, a crowd sensing privacy protection and incentive method called tangle network incentive policy (TNIP) based on the tangle network was proposed. Firstly, guided by the concept of privacy protection, the sensing network adopted the tangle network framework, which ensured network security through a distributed recording approach. Secondly, to address issues such as malicious competition and the proof of work threshold introduced by blockchain, a PoW (proof of work) control algorithm was designed to lower the participation threshold and mitigate malicious competition. Furthermore, the privacy of sensing data was strengthened by utilizing the ECDSA (elliptic curve digital signature algorithm) based digital signature method, enhancing the security of the data. Subsequently, a quantitative evaluation model was used to assess the risks of privacy leakage, and a trusted milestone transaction was employed to control network risks and enhance the privacy of the sensing network. Finally, by utilizing real-world datasets and conducting simulation experiments, a comparative analysis of the incentive effect and privacy protection methods proposed by TNIP was performed. The experimental results demonstrate that the TNIP method achieves an average participation rate improvement of 8.79% and 5.93% over the CSII (cross-space multi-interaction-based-dynamic incentive scheme) method and TCS (tangle-net crowd sensing) method, respectively, for the three categories of tasks.