A Hybrid Trajectory Prediction Framework for Automated Vehicles with Attention Mechanisms

Mingqiang Wang,Lei Zhang,Jun Chen,Zhiqiang Zhang,Zhenpo Wang,Dongpu Cao
DOI: https://doi.org/10.1109/tte.2023.3346668
IF: 6.519
2024-01-01
IEEE Transactions on Transportation Electrification
Abstract:The driving safety of automated vehicles is largely dependent on accurately predicting the motions of surrounding vehicles. However, the existing approaches ignore the impact of the ego vehicle’s future behaviors on the surrounding vehicles and lack model explainability for the prediction results. To tackle this issue, a hybrid trajectory prediction framework based on Long Short-Term Memory (LSTM) encoding is proposed. It introduces a reactive social convolution structure to model the planned trajectory of the ego vehicle with the historical trajectories of the surrounding vehicles to reduce uncertainty in potential trajectories. Furthermore, a spatio-temporal attention mechanism is presented to quantitatively describe the contributions of historical trajectories and interactions among the surrounding vehicles to the prediction results by appropriate weights setting. Finally, the proposed scheme is comprehensively evaluated based on the NGSIM and HighD datasets. The results demonstrate that the proposed approach can elucidate the prediction process from a spatio-temporal perspective and outperforms other state-of-the-art methods under different scenarios. The Root-Mean-Square errors in the NGSIM and HighD datasets are reduced to less than 3.65 m and 2.36 m over a time horizon of 5 s , respectively. The qualitative analysis on the reliability and reactivity are also presented.
What problem does this paper attempt to address?