AN ADAPTIVE SMARTPHONE HYBRID INDOOR POSITIONING SOLUTION INCORPORATING HETEROGENEOUS SENSORS
Keywords: Indoor Positioning, Smartphone Navigation, Pedestrian Localization, Multi-sensor Integration
Abstract. Indoor positioning algorithms based on Wi-Fi or Bluetooth low energy (BLE) have been widely used for indoor navigation and Internet of Things (IoT) applications, due to their low cost. However, their positioning accuracy is between 2–5 meters, which is insufficient for many emerging applications requiring higher accuracy. New high-precision positioning technologies, such as light, acoustic, and Bluetooth Angle of Arrival (AoA)-based positioning systems, have been developed, and they can achieve decimeter-level positioning accuracy. However, these high-precision positioning technologies suffer from high costs. In this paper, an adaptive smartphone hybrid indoor positioning solution incorporating heterogeneous sensors is proposed to achieve high accuracy and reliable localization. Multiple indoor scenes are identified and the quality of pedestrian movement is detected, thus enabling the adaptive estimation and adjustment of measurement noise and process noise in the filtering procedure. High-precision sensors (i.e. Bluetooth array, acoustic sensors and light sensors) are utilized in the proposed solution as controlling anchors to improve the positioning accuracy. Low-cost sensors (i.e. Wi-Fi, BLE and inertial measurement unit (IMU) sensors), are integrated to achieve wide-area coverage. An autonomous outlier detection method is developed to improve the positioning accuracy and reliability. Experimental results show that the proposed solution can achieve accuracy of 0.98 m in the test scenario.