RSSI-Based Indoor Localization Using Machine Learning for Wireless Sensor Networks: A Recent Review

The significance of device localization has increased in response to the current challenges of indoor positioning systems. This need is particularly severe in wireless sensor networks, where node awareness of its location within the sensor field is advantageous for tasks such as event monitoring. Numerous review papers have categorized indoor localization algorithms using conventional machine learning methods. However, limited attention has been devoted to works incorporating alternative classification schemes and considering learning-based localization systems, as proposed in this study. This paper introduces a review of indoor localization machine learning approaches to extend the discussion by focusing on learning models that dynamically update with each new data point received, categorizing approaches into fingerprinting and trilateration techniques from received signal strength indicators. Thus, the proposed study lays the foundation for future research endeavors in indoor localization, especially those integrating incremental learning and continuously updating models with new data.