Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

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

Haicheng Liao, Zhenning Li, Chengyue Wang, Bonan Wang, Hanlin Kong, Yanchen Guan, Guofa Li, Zhiyong Cui, Chengzhong Xu

TL;DR

This work tackles trajectory prediction in mixed autonomy by integrating cognitive notions of perceived safety into a predictive framework. It introduces a three-module architecture—Perceived Safety-Aware, Priority-Aware, and Interaction-Aware with a Leanformer backbone and a multimodal Gaussian decoder—that jointly model safety, behavior, and social interactions. The approach leverages Quantitative Safety Assessment metrics and Driver Behavior Profiling to capture human-like decision making, achieving state-of-the-art gains on NGSIM, HighD, and MoCAD, and demonstrating robustness to missing data and limited training data. The results indicate strong practical potential for safer and more efficient autonomous driving in real-world, data-challenged environments.

Abstract

As autonomous driving technology progresses, the need for precise trajectory prediction models becomes paramount. This paper introduces an innovative model that infuses cognitive insights into trajectory prediction, focusing on perceived safety and dynamic decision-making. Distinct from traditional approaches, our model excels in analyzing interactions and behavior patterns in mixed autonomy traffic scenarios. It represents a significant leap forward, achieving marked performance improvements on several key datasets. Specifically, it surpasses existing benchmarks with gains of 16.2% on the Next Generation Simulation (NGSIM), 27.4% on the Highway Drone (HighD), and 19.8% on the Macao Connected Autonomous Driving (MoCAD) dataset. Our proposed model shows exceptional proficiency in handling corner cases, essential for real-world applications. Moreover, its robustness is evident in scenarios with missing or limited data, outperforming most of the state-of-the-art baselines. This adaptability and resilience position our model as a viable tool for real-world autonomous driving systems, heralding a new standard in vehicle trajectory prediction for enhanced safety and efficiency.

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

TL;DR

This work tackles trajectory prediction in mixed autonomy by integrating cognitive notions of perceived safety into a predictive framework. It introduces a three-module architecture—Perceived Safety-Aware, Priority-Aware, and Interaction-Aware with a Leanformer backbone and a multimodal Gaussian decoder—that jointly model safety, behavior, and social interactions. The approach leverages Quantitative Safety Assessment metrics and Driver Behavior Profiling to capture human-like decision making, achieving state-of-the-art gains on NGSIM, HighD, and MoCAD, and demonstrating robustness to missing data and limited training data. The results indicate strong practical potential for safer and more efficient autonomous driving in real-world, data-challenged environments.

Abstract

As autonomous driving technology progresses, the need for precise trajectory prediction models becomes paramount. This paper introduces an innovative model that infuses cognitive insights into trajectory prediction, focusing on perceived safety and dynamic decision-making. Distinct from traditional approaches, our model excels in analyzing interactions and behavior patterns in mixed autonomy traffic scenarios. It represents a significant leap forward, achieving marked performance improvements on several key datasets. Specifically, it surpasses existing benchmarks with gains of 16.2% on the Next Generation Simulation (NGSIM), 27.4% on the Highway Drone (HighD), and 19.8% on the Macao Connected Autonomous Driving (MoCAD) dataset. Our proposed model shows exceptional proficiency in handling corner cases, essential for real-world applications. Moreover, its robustness is evident in scenarios with missing or limited data, outperforming most of the state-of-the-art baselines. This adaptability and resilience position our model as a viable tool for real-world autonomous driving systems, heralding a new standard in vehicle trajectory prediction for enhanced safety and efficiency.
Paper Structure (17 sections, 30 equations, 3 figures, 4 tables)

This paper contains 17 sections, 30 equations, 3 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Problem Formulation
  4. Trajectory Prediction Model
  5. Perceived Safety Aware Module
  6. Priority-Aware Module
  7. Interaction-Aware Module
  8. Multimodal Decoder
  9. Experiments
  10. Experimental Setup
  11. Experimental Results
  12. Performance Comparison on Complete Test Set
  13. Performance Comparison on Missing Test Set
  14. Performance on a Limited 25% Training Set
  15. Qualitative Results
  16. ...and 2 more sections

Figures (3)

  • Figure 1: Perceived safety and its influence on decisions of drivers with different driving behaviors.
  • Figure 2: Architecture of proposed trajectory prediction model.
  • Figure 3: Visual insights from the NGSIM dataset depict two complex driving scenarios: (a) driving straight and (b) merging to the right. The target vehicle is marked by red triangles, with surrounding vehicles represented by other triangles. Subfigures (a-1) and (b-1) present the outcomes of driver behavior profiling, whereas (a-2) and (b-2) illustrate the perceived safety metrics for each vehicle alongside their predicted trajectories. A vehicle shaded in deeper brown signals potential aggressive behavior, while a darker blue shade indicates a heightened risk to the target vehicle and vice versa.