Abstract:Accurately predicting the trajectory of surrounding vehicles is a critical challenge for autonomous vehicles. In complex traffic scenarios, there are two significant issues with the current autonomous driving system: the cognitive uncertainty of prediction and the lack of risk awareness, which limit the further development of autonomous driving. To address this challenge, we introduce a novel trajectory prediction model that incorporates insights and principles from driving behavior, ethical decision-making, and risk assessment. Based on joint prediction, our model consists of interaction, intention, and risk assessment modules. The dynamic variation of interaction between vehicles can be comprehensively captured at each timestamp in the interaction module. Based on interaction information, our model considers primary intentions for vehicles to enhance the diversity of trajectory generation. The optimization of predicted trajectories follows the advanced risk-aware decision-making principles. Experimental results are evaluated on the DeepAccident dataset; our approach shows its remarkable prediction performance on normal and accident scenarios and outperforms the state-of-the-art algorithms by at least 28.9\% and 26.5\%, respectively. The proposed model improves the proficiency and adaptability of trajectory prediction in complex traffic scenarios. The code for the proposed model is available at <a class="link-external link-https" href="https://sites.google.com/view/ir-prediction" rel="external noopener nofollow">this https URL</a>.
What problem does this paper attempt to address?
### Problems the Paper Aims to Solve
This paper aims to address the critical challenge of accurately predicting the trajectories of surrounding vehicles in complex traffic scenarios for autonomous driving vehicles. Specifically, current autonomous driving systems face two significant issues in complex traffic environments:
1. **Cognitive Uncertainty in Prediction**: Due to the randomness and uncertainty of driver behavior, existing autonomous driving systems struggle to accurately predict the behavior of surrounding vehicles.
2. **Lack of Risk Awareness**: Existing autonomous driving systems often lack the assessment of potential risks in decision-making, which limits their performance in high-risk scenarios.
To tackle these challenges, the authors propose a new trajectory prediction model that integrates principles of driving behavior, ethical decision-making, and risk assessment. The model consists of an interaction module, an intention module, and a risk assessment module, which can comprehensively capture the dynamic interactions between vehicles and generate diverse trajectories based on intentions. Additionally, the model optimizes the predicted trajectories through advanced risk-aware decision-making principles, thereby improving prediction performance in both normal and accident scenarios.
### Main Contributions of the Model
1. **Intention Feature Module**: By focusing on intention-related trajectories rather than all possible motion patterns, the accuracy and efficiency of trajectory prediction are improved.
2. **Trajectory Optimization Module**: By considering potential risks, the vehicle can make more adaptive and safer decisions.
3. **Performance Improvement**: In both normal and accident scenarios, the model outperforms existing state-of-the-art algorithms by at least 28.9% and 26.5%, respectively, demonstrating its effectiveness and adaptability in complex traffic environments.
### Experimental Results
The experimental results were evaluated on the DeepAccident dataset, showing that the model performs excellently in both normal and accident scenarios. Particularly in accident scenarios, its performance significantly surpasses other benchmark models. This indicates that the model has better navigation and prediction capabilities in high-risk situations.
### Conclusion
This paper proposes a joint prediction model that combines interaction, intention, and risk assessment, effectively solving the trajectory prediction problem for autonomous driving systems in complex traffic scenarios. Through detailed experiments and ablation studies, the importance of each module is validated, emphasizing the necessity of incorporating traffic behavior principles into the prediction system. In the future, the authors plan to further study the planning module corresponding to this prediction module to build a fully functional autonomous driving system.