To address the challenge of fluctuations and mutations in received signal strength indication (RSSI) caused by environmental influences in WiFi-based indoor positioning applications, which ultimately leads to reduced accuracy, we propose a WiFi hybrid filtering RSSI indoor positioning algorithm.Firstly, an adaptive threshold filter is developed based on the WiFi signal distance attenuation model to effectively eliminate outliers from the observed signal.Next, a hybrid filtering algorithm (ATKM) was designed based on adaptive threshold filtering, Kalman filtering, and Mean filtering to filter RSSI data.Finally, we propose a weighted k-nearest neighbor algorithm (SEWKNN) for position estimation, which utilizes Spearman-Euclidean distance.Experimental results conducted in real-world environments demonstrate that the ATKM filtering algorithm significantly reduces the fluctuations in RSSI signals.The SEWKNN algorithm achieves an average positioning error of 1.17 m in indoor environments and 1.53 m in corridor environments.Compared to the traditional WKNN algorithm, the proposed SEWKNN algorithm reduces the average positioning error by 18.18％ and 16.84％ respectively, indicating its superior performance.
关键词
接收信号强度指示/室内定位/混合滤波/指纹匹配/斯皮尔曼相关系数
Key words
received signal strength indication/Indoor positioning/Hybrid filter/Fingerprint matching/Spearman's rank correlation coefficient
维普
万方数据