Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

EPN: An Ego Vehicle Planning-Informed Network for Target Trajectory Prediction

Saiqian Peng, Duanfeng Chu, Guanjie Li, Liping Lu, Jinxiang Wang

TL;DR

EPN addresses the problem of multimodal target trajectory prediction in complex traffic by incorporating the ego vehicle's planned future trajectory as an input to model mutual influence, and by leveraging a CVAE-based endpoint predictor with a correction mechanism to resolve intention ambiguity. The architecture fuses spatiotemporal features through a feature fusion encoder and a social pooling module, then predicts multiple plausible endpoints and generates complete trajectories via an LSTM decoder conditioned on corrected endpoints and relative displacements. Empirical results on NGSIM and HighD show substantial reductions in RMSE, ADE, and FDE compared to state-of-the-art methods, underscoring the value of ego planning and endpoint-based predictions for robust trajectory forecasting. The approach offers practical impact for safer autonomous driving by more accurately anticipating nearby vehicle behavior under dynamic, multimodal conditions.

Abstract

Trajectory prediction plays a crucial role in improving the safety of autonomous vehicles. However, due to the highly dynamic and multimodal nature of the task, accurately predicting the future trajectory of a target vehicle remains a significant challenge. To address this challenge, we propose an Ego vehicle Planning-informed Network (EPN) for multimodal trajectory prediction. In real-world driving, the future trajectory of a vehicle is influenced not only by its own historical trajectory, but also by the behavior of other vehicles. So, we incorporate the future planned trajectory of the ego vehicle as an additional input to simulate the mutual influence between vehicles. Furthermore, to tackle the challenges of intention ambiguity and large prediction errors often encountered in methods based on driving intentions, we propose an endpoint prediction module for the target vehicle. This module predicts the target vehicle endpoints, refines them using a correction mechanism, and generates a multimodal predicted trajectory. Experimental results demonstrate that EPN achieves an average reduction of 34.9%, 30.7%, and 30.4% in RMSE, ADE, and FDE on the NGSIM dataset, and an average reduction of 64.6%, 64.5%, and 64.3% in RMSE, ADE, and FDE on the HighD dataset. The code will be open sourced after the letter is accepted.

EPN: An Ego Vehicle Planning-Informed Network for Target Trajectory Prediction

TL;DR

EPN addresses the problem of multimodal target trajectory prediction in complex traffic by incorporating the ego vehicle's planned future trajectory as an input to model mutual influence, and by leveraging a CVAE-based endpoint predictor with a correction mechanism to resolve intention ambiguity. The architecture fuses spatiotemporal features through a feature fusion encoder and a social pooling module, then predicts multiple plausible endpoints and generates complete trajectories via an LSTM decoder conditioned on corrected endpoints and relative displacements. Empirical results on NGSIM and HighD show substantial reductions in RMSE, ADE, and FDE compared to state-of-the-art methods, underscoring the value of ego planning and endpoint-based predictions for robust trajectory forecasting. The approach offers practical impact for safer autonomous driving by more accurately anticipating nearby vehicle behavior under dynamic, multimodal conditions.

Abstract

Trajectory prediction plays a crucial role in improving the safety of autonomous vehicles. However, due to the highly dynamic and multimodal nature of the task, accurately predicting the future trajectory of a target vehicle remains a significant challenge. To address this challenge, we propose an Ego vehicle Planning-informed Network (EPN) for multimodal trajectory prediction. In real-world driving, the future trajectory of a vehicle is influenced not only by its own historical trajectory, but also by the behavior of other vehicles. So, we incorporate the future planned trajectory of the ego vehicle as an additional input to simulate the mutual influence between vehicles. Furthermore, to tackle the challenges of intention ambiguity and large prediction errors often encountered in methods based on driving intentions, we propose an endpoint prediction module for the target vehicle. This module predicts the target vehicle endpoints, refines them using a correction mechanism, and generates a multimodal predicted trajectory. Experimental results demonstrate that EPN achieves an average reduction of 34.9%, 30.7%, and 30.4% in RMSE, ADE, and FDE on the NGSIM dataset, and an average reduction of 64.6%, 64.5%, and 64.3% in RMSE, ADE, and FDE on the HighD dataset. The code will be open sourced after the letter is accepted.

Paper Structure

This paper contains 20 sections, 30 equations, 3 figures, 4 tables.

Table of Contents

  1. INTRODUCTION
  2. RELATED WORK
  3. Physics-based methods
  4. Machine learning methods
  5. Deep learning methods
  6. PROBLEM FORMULATION
  7. Model Input
  8. Model Output
  9. MODEL ARCHITECTURE
  10. Feature Fusion Encoding
  11. Target's Endpoint Prediction
  12. LSTM Decoder
  13. Loss Function
  14. EXPERIMENTS
  15. Dataset and Data Preprocessing
  16. ...and 5 more sections

Figures (3)

  • Figure 1: Vehicle classification and region of interest division in a driving scenario.
  • Figure 2: Overall architecture of EPN model.
  • Figure 3: Visualization results in left-turn, straight-ahead and right-turn driving scenarios.