【摘要】 由于广播式自动相关监视（automatic dependent surveillance-broadcast, ADS-B）报文格式完全公开且重要数据缺乏加密措施，因此敏感定位数据面临泄露的风险。为解决不可信第三方基于位置服务（untrusted third-party location based service, UT-LBS）在收集和发布ADS-B报文定位数据时造成的隐私泄露问题，提出一种阶段式等差随机响应（phased equal-step randomized response, PERR）隐私保护方法。首先，对三维定位数据进行降维处理，通过二进制交叉编码生成一维八进制定位信息编码。其次，使用阶段式等差随机响应方法对一维定位信息编码进行扰动，通过所提出的等差扰动的思想对随机响应空间进行阶梯式分组。然后，将随机响应分成基于等差隐私的随机响应和基于等差概率的自适应随机响应空间划分两个阶段，提高随机响应过程中数据的隐私性。最后，通过所提出的基于阶段式概率加权的卡尔曼滤波器估算方法，对随机响应后的隐私数据进行频率估算，并分析证明该方法的隐私性和可行性。实验结果表明，所提方法在实现ADS-B敏感定位数据隐私保护的同时，有效减小隐私保护后数据的频数统计与原始数据的频数统计之间的误差，保证了数据的有效性。更多还原

【Abstract】 Due to the completely open format of automatic dependent surveillance-broadcast(ADS-B) messages and the lack of encryption for critical data, sensitive location information was put at risk of exposure. To address the privacy leakage issues caused by untrusted third-party Location-Based Services during the collection and publication of ADS-B message location data, a phased equal-step randomized response privacy protection method was proposed. Dimensionality reduction was performed on the three-dimensional location data, and a one-dimensional octal location information code was generated through binary crossover encoding. A phased equal difference randomized response method was employed to perturb the one-dimensional location information code. The concept of equal difference perturbation was utilized to create ladder-like groupings in the randomized response space. Subsequently, the randomized response process was divided into two phases: a random response based on equal difference privacy, and an adaptive random response space partitioned by equal difference probability, thereby enhancing data privacy throughout the randomization process. Finally, a phase-based probabilistic weighted Kalman filter estimation method was proposed to estimate the frequency of the privacy data after randomization. The privacy and feasibility of this approach were demonstrated through theoretical analysis. Experimental results indicate that the proposed method effectively protects the privacy of ADS-B sensitive location data while significantly reducing the error between the frequency statistics of the privacy-protected data and the original data, thus ensuring data validity.更多还原