Pivot-Centric Trajectory Prediction: Bridging Long Horizons via Dynamical Guidance

2026-08-04

AI Summary

Long-horizon trajectory prediction is prone to performance degradation due to error accumulation and the absence of intermediate guidance. This work proposes a Pivot-Centric Prediction framework that decomposes long-term forecasting into multi-scale short-term subtasks by introducing "pivot points." The approach operates in two stages: first predicting key pivot points, then refining the full trajectory conditioned on these pivots. By integrating global map context, agent interactions, and local details, the method provides dynamic intermediate supervision to mitigate error propagation and can be flexibly incorporated into existing architectures. Evaluated on Argoverse I and II, the framework significantly enhances the performance of state-of-the-art methods; when integrated with QCNet, it surpasses all published non-ensemble approaches on the Argoverse II leaderboard.

Abstract

Forecasting precise future motion of surrounding agents is essential for reliable autonomous vehicles. However, as the demand for longer prediction horizons increases, existing endpoint-completion or iterative-refine methods increasingly struggle with weak guidance and compounding errors. To tackle the long-horizon prediction challenge, we propose Pivot-Centric Trajectory Prediction (PCTP). By introducing ``pivots'' and focusing on predicting pivot points along extended trajectories, we divide the long-term prediction task into short-term sub-tasks at various scales. Specifically, PCTP decouples the long-term trajectory predicting process into two processes: pivot prediction and pivot-based trajectory refinement. The pivot prediction process aims to utilize global map context and agent-to-agent interactions to identify these ``pivot points'', while the pivot-based trajectory refinement process focuses on local map details and refines the short-term trajectory based on predicted ``pivot points''. Compared with existing methods, PCTP provides more intermediate guidance while reducing compounding errors. Moreover, PCTP is a flexible approach that can be integrated into most state-of-the-art trajectory prediction models. Experimental results show that PCTP improves the prediction accuracy of leading models on both Argoverse I and Argoverse II datasets with minimal impact on model size. Specifically, PCTP combined with QCNet outperforms all published ensemble-free methods on the Argoverse II leaderboard at submission.

Problem

• trajectory prediction
• long-horizon forecasting
• autonomous driving
• compounding errors
• motion forecasting

Innovation

• Pivot-Centric Trajectory Prediction
• long-horizon prediction
• pivot points
• trajectory refinement
• compounding error reduction