| Passive RFID tags are widely used for indoor localization. However, a variety of environmental factors tend to reduce the localization accuracy. This research explores how different relative dielectric constants impact RSSI-based RFID Indoor localization. Three ML algorithms (1) k-NN, (2) XGBoost, and (3) Decision Tree are used for the indoor localization of Cantaloupe, Cabbage, and Pineapple which have different dielectric values. XGBoost achieved an accuracy of 54.3% for cantaloupe which has a low dielectric value. For pineapple and cabbage which has relatively close dielectric value, the same algorithm achieved comparable accuracy of 76.9% and 78.34% respectively. This research demonstrates the importance of the object’s dielectric constant when developing RSSI-based indoor localization systems. |