Abstract:Trajectory prediction is one of the key components of the autonomous driving software stack. Accurate prediction for the future movement of surrounding traffic participants is an important prerequisite for ensuring the driving efficiency and safety of intelligent vehicles. Trajectory prediction algorithms based on artificial intelligence have been widely studied and applied in recent years and have achieved remarkable results. However, complex artificial intelligence models are uncertain and difficult to explain, so they may face unintended failures when applied in the real world. In this paper, a self-aware trajectory prediction method is proposed. By introducing a self-awareness module and a two-stage training process, the original trajectory prediction module's performance is estimated online, to facilitate the system to deal with the possible scenario of insufficient prediction function in time, and create conditions for the realization of safe and reliable autonomous driving. Comprehensive experiments and analysis are performed, and the proposed method performed well in terms of self-awareness, memory footprint, and real-time performance, showing that it may serve as a promising paradigm for safe autonomous driving.

What problem does this paper attempt to address?

The problem this paper attempts to address is that in autonomous driving systems, existing AI-based trajectory prediction models, despite achieving significant results in prediction performance, have uncertainties, are difficult to interpret, and may face unexpected failures in practical applications. Specifically, these models may lead to unacceptable decision errors when dealing with unknown long-tail scenarios, potentially causing severe traffic accidents. To solve this problem, the paper proposes a self-aware trajectory prediction method. By introducing a self-aware module and a two-stage training process, this method can estimate the performance of the original trajectory prediction module during online operation, thereby helping the system to timely respond to potential prediction deficiencies, creating conditions for safe and reliable autonomous driving. The main contributions include: 1. **Novel self-aware trajectory prediction framework**: A new, computationally efficient self-aware module is introduced, which can diagnose prediction results by obtaining real-time information from the intermediate and output layers of the original trajectory prediction module. 2. **New training and inference process**: A two-stage training process is proposed to optimize the self-aware module without affecting the performance of the original trajectory prediction module. The training of the self-aware module does not require modifying the parameters of the already trained trajectory prediction module, only obtaining part of its information, thus having strong flexibility and practicality. 3. **Comprehensive experiments and analysis**: By comparing with various baseline methods, the proposed method's superiority in self-awareness capability, memory usage, and real-time performance is demonstrated. Additionally, detailed ablation studies and quantitative and qualitative analyses provide in-depth exploration. In summary, this paper aims to improve the safety, reliability, and robustness of autonomous driving systems by endowing trajectory prediction models with self-awareness capabilities.

Self-Aware Trajectory Prediction for Safe Autonomous Driving

Intention-based and Risk-Aware Trajectory Prediction for Autonomous Driving in Complex Traffic Scenarios

A Cognitive-Based Trajectory Prediction Approach for Autonomous Driving

A Cognitive-Driven Trajectory Prediction Model for Autonomous Driving in Mixed Autonomy Environment

Enhanced Multimodal Trajectory Prediction for Autonomous Vehicles Using Advanced Diffusion Model Techniques

MFTraj: Map-Free, Behavior-Driven Trajectory Prediction for Autonomous Driving

An Integrating Comprehensive Trajectory Prediction with Risk Potential Field Method for Autonomous Driving

AI-TP: Attention-based Interaction-aware Trajectory Prediction for Autonomous Driving

Risk-Aware Vehicle Trajectory Prediction Under Safety-Critical Scenarios

Characterized Diffusion and Spatial-Temporal Interaction Network for Trajectory Prediction in Autonomous Driving

Long-Tail Prediction Uncertainty Aware Trajectory Planning for Self-driving Vehicles.

A Probabilistic Architecture of Long-Term Vehicle Trajectory Prediction for Autonomous Driving

A Cognition‐inspired Trajectory Prediction Method for Vehicles in Interactive Scenarios

Advancing Autonomous Driving Safety Through LLM Enhanced Trajectory Prediction

Probabilistic Trajectory Prediction in Intelligent Driving

Trajectory Planning and Safety Assessment of Autonomous Vehicles Based on Motion Prediction and Model Predictive Control.

Multimodal Trajectory Predictions for Autonomous Driving using Deep Convolutional Networks

Less is More: Efficient Brain-Inspired Learning for Autonomous Driving Trajectory Prediction

A Hybrid Trajectory Prediction Framework for Automated Vehicles with Attention Mechanisms

Probabilistic Long-term Vehicle Trajectory Prediction Via Driver Awareness Model

What Truly Matters in Trajectory Prediction for Autonomous Driving?