Interrupted Track Association Based on CNN-BiLSTM Bidirectional Prediction and Fuzzy Analysis

With the advancement of modern multi-sensor tracking technology, track association plays a critical role in applications such as target tracking and air traffic control. To address the issue of cross-regional track interruptions in multi-radar cooperative systems, this paper proposes an interrupted track association method based on CNN-BiLSTM bidirectional prediction and fuzzy analysis. The method utilizes Convolutional Neural Networks (CNNs) to extract spatial features of track data, and employs Bidirectional Long Short-Term Memory (BiLSTM) networks to perform temporal modeling by predicting the future trajectory of the target forward in time and inferring its historical trajectory backward in time. This approach enables more accurate prediction and recovery of the target's trajectory during interruption periods. In addition, a fuzzy logic analysis mechanism is introduced to enhance the robustness of the algorithm in handling uncertainty by establishing fuzzy logic rules to quantify ambiguity in data association. Experimental results demonstrate that the proposed method effectively reduces the probabilities of missed associations and false associations.