Research on Indoor Localization Algorithm for Emergency Rescue Personnel Based on INS/UWB

To address the problem of low rescuer positioning accuracy in complex indoor rescue scenarios, a tightly combined positioning algorithm based on an adaptive weighting inertial navigation system/ultra-wideband (INS/UWB) is proposed. The algorithm innovatively integrates deep learning and error-resistant estimation theory and reduces the positioning error through multi-dimensional optimization. A zero velocity detection algorithm is developed based on a hybrid model of convolutional neural networks (CNNs), bidirectional long short-term memory (BiLSTM), and an attention mechanism, which can accurately detect the zero velocity interval (ZVI) of a person's gait. The Zero Velocity Interval (ZVI) of a person's gait can be accurately recognized with a recognition accuracy of 98.2%, which is 15.7% higher than that of the traditional method. For the indoor UWB non-line-of-sight (NLOS) error, the IGG III error-resistant model is introduced to dynamically adjust the weights of the UWB range values and adaptively adjust the configuration of the system state covariance matrix and the measurement matrix. The experimental results show that under the indoor mixed occlusion conditions, the positioning accuracy is improved by 16.3% compared with the standard EKF algorithm, and the positioning error is controlled within 0.89m, indicating that the adaptive weighting INS/UWB tight combination positioning algorithm proposed in this paper is able to effectively use the IMU and UWB information to achieve a high-precision positioning function in indoor rescue scenarios and meet the positioning requirements in emergency rescue scenarios.