SmartFPS: Neural Network based Wireless-inertial fusion positioning system

The current fusion positioning systems are mainly based on filtering algorithms, such as Kalman filtering or particle filtering. However, the system complexity of practical application scenarios is often very high, such as no ise modeling in pedestrian inertial navigation systems, or environmental noise modeling in fingerprint matching and localization algorithms. To solve this problem, this paper proposes a fusion positioning system based on deep learning and proposes a transf er learning strategy for improving the performance of neural network models for samples with different distributions. The results show that in the whole floor scenario, the average positioning accuracy of the fusion network is 0.506 m . The experiment results of transfer learning show that the estimation accuracy of the inertial navigation positioning step size and rotation angle of different pedestrians can be improved by 53.3% on average, the Bluetooth positioning accuracy of different devices can be improved by 33.4%, and the fusion can be improved by 31.6%. However, the current mature GPS positioning is usually unable to locate effectively indoors due to irregular attenuati o n caused b y the occlusion of GPS signals by clouds, building walls, and ceilings. Because of this, the new technology of indoor positioning system was proposed. Each positioning technique has its advantages as well as its limitations. For example, inertial navigation positioning is prone to accumulative errors d u e to system noise and drift [19, 20]; positioning signals such as Wi - Fi and Bluetooth fluctuate and signal attenuation is difficult to model, so traditional methods such as trilateral positioning [21, 22] are directly used for positioning accuracy. Manuscript received xx, xx; revised x x, xx; accepted xx, xx . Luchi Hua and Jun Yang are with Southeast University, 2 Sipailou, Nanjing 210096, China (e - mail: 1 046902506@qq.com, Jun Yang is corres ponding author. In general, high - precision, and high - stable positioning performance cannot be obtained based on a single positioning system. On the other hand, high - cost positioning systems canno t be used for civilian use. Especially in the pedestrian positioning scenario, factors such as portability and cost need to be considered. Therefore, most positioning solutions obtain multiple sensor data from the user's mobile terminal to co o rdinate posit ioning services, such as the gyroscope and accelerometer of the inertial unit in the smartphone. Wi - Fi and Bluetooth modules, etc. Due to the limitations of various positioning technologies, fusion positioning systems have been widely studied in recent year s, such as Wi - Fi, Bluetooth, Lidar, and inertial navigation fusion [27 - 32]. Simply put, data fusion is the process of combining data from multiple sensors and related information to achieve more specific inferences than can be achieved with a single sensor.