Research on Motion Target Positioning Using Smartphone Inertial Measurement Units

This study has developed a motion target positioning method leveraging the Inertial Measurement Unit (IMU) data embedded in smartphones, addressing the challenges of indoor positioning services such as complex deployment, high costs, and limited accessibility. The research integrates an attention mechanism with a Long Short-Term Memory (LSTM) model to analyze the relationship between the accelerometer and gyroscope data from smartphones and the velocity of moving targets, enabling precise regression and prediction of motion trajectories. By incorporating the attention mechanism into the RoNIN network model, the model's capability to process complex and nonlinear data is significantly enhanced, thereby improving the accuracy and generalization of predictions. The study designed an enhanced LSTM module with integrated attention mechanisms within the RoNIN algorithm, which includes a position embedding layer, feature extraction layer, and trajectory output layer. This design not only optimizes the feature extraction process but also enhances the capture of contextual features related to pedestrian movement. The conclusion of this paper provides a detailed account of the evaluation results, demonstrating the model's superiority in time series prediction tasks.