GPS real-time and long-term tracking algorithm for mobile radioactive sources

Time error accumulation and large fluctuations in position data accuracy under different environments and motion states result in low tracking stability. Therefore, a real-time, long-term GPS tracking algorithm for mobile radioactive sources is proposed. A sliding average filter technique is used to weighted average and correct the longitude and latitude coordinates of the raw GPS data. GPS discipline is then used to calculate frequency accuracy to synchronize the time base, thereby improving position data accuracy and providing reliable temporal and spatial information for subsequent tracking. Data from auxiliary sensors such as the IMU and barometer are integrated, and reliable data is selected using a consistency metric. An information entropy weighting strategy is employed to dynamically balance the confidence of multi-source data. The fusion results are then fed back to the sliding average filter module for iterative optimization, ensuring stable and accurate tracking of the mobile radioactive source even in complex environments. Experiments show that this method reduces the average positioning drift rate to 1.53 times per hour, maintains a low and stable positioning loss rate across all scenarios, and significantly outperforms comparative methods in interference environments, with scatter points highly concentrated around the ideal area. Experiments demonstrate that this method significantly improves tracking stability through multi-source sensor data fusion, adaptive filtering, and tracking strategy adjustment.