Integrated BLE PIFA Sensor With Machine Learning for Liquid Level Detection in IoT Applications

This article presents an integrated Bluetooth low energy (BLE) planar inverted-F antenna (PIFA) sensor designed for noncontact liquid-level detection using microwave sensing in the 2.4 GHz ISM band. The bottom and side placements of the PIFA are investigated, revealing distinct resonant frequency behavior in response to liquid-level changes. While the side PIFA exhibits a near-linear response within a specific range, the bottom PIFA shows nonlinear due to near-field interference and mode competition. These nonlinearities are addressed by employing a machine learning approach called random forest regression (RFR), achieving high-precision volume prediction with an R2 of 0.9885. The combination of microwave sensing and machine learning demonstrates a robust, compact, and scalable solution for accurate liquid-level monitoring in Internet of Things (IoT) applications. The proposed system is further integrated with BLE in IoT applications, enabling compact, low cost, real-time remote monitoring, and alerting.