The paper presents a novel trajectory prediction model, GaVa, designed for autonomous vehicles (AVs) in mixed autonomy environments. The model integrates principles of human cognition and observational behavior to enhance the accuracy and adaptability of trajectory predictions. Key contributions include: 1. **Adaptive Visual Sector Mechanism**: This mechanism dynamically adjusts the visual field based on factors like speed, spatial orientation, and proximity, mimicking human drivers' attention allocation. 2. **Dynamic Traffic Graph**: Utilizing Convolutional Neural Networks (CNN) and Graph Attention Networks (GAT), this mechanism captures spatio-temporal dependencies among agents, improving the model's ability to predict trajectories. 3. **Model Performance**: Benchmark tests on NGSIM, HighD, and MoCAD datasets show that GaVa outperforms state-of-the-art baselines by at least 15.2%, 19.4%, and 12.0%, respectively. The research highlights the potential of integrating human cognition principles into trajectory prediction algorithms, making AVs more proficient and adaptable in complex traffic scenarios. The code for the proposed model is available on GitHub.The paper presents a novel trajectory prediction model, GaVa, designed for autonomous vehicles (AVs) in mixed autonomy environments. The model integrates principles of human cognition and observational behavior to enhance the accuracy and adaptability of trajectory predictions. Key contributions include: 1. **Adaptive Visual Sector Mechanism**: This mechanism dynamically adjusts the visual field based on factors like speed, spatial orientation, and proximity, mimicking human drivers' attention allocation. 2. **Dynamic Traffic Graph**: Utilizing Convolutional Neural Networks (CNN) and Graph Attention Networks (GAT), this mechanism captures spatio-temporal dependencies among agents, improving the model's ability to predict trajectories. 3. **Model Performance**: Benchmark tests on NGSIM, HighD, and MoCAD datasets show that GaVa outperforms state-of-the-art baselines by at least 15.2%, 19.4%, and 12.0%, respectively. The research highlights the potential of integrating human cognition principles into trajectory prediction algorithms, making AVs more proficient and adaptable in complex traffic scenarios. The code for the proposed model is available on GitHub.