Inferring Missing Trajectory Data with Temporal Convolutional Networks

2026-07-27

AI Summary

This study addresses the challenge of reconstructing continuous and smooth trajectory segments in real-world scenarios, where missing data commonly arise due to sensor failures, communication disruptions, or occlusions. To overcome the limitations of conventional approaches that rely solely on historical context, this work proposes a non-causal Symmetric Dilated Temporal Convolutional Network (Symmetric Dilated TCN) that leverages contextual information from both past and future time steps around the missing segment. The model employs a composite loss function integrating weighted mean squared error, boundary continuity penalties, and a smoothness regularizer to enhance reconstruction fidelity. Evaluated on a 2D trajectory dataset with 20% randomly masked entries, the proposed method demonstrates superior performance across multiple metrics—including R², MSE, and MAE—confirming its effectiveness and robustness in trajectory completion tasks.

Abstract

Trajectory data collected in real-world settings is frequently incomplete due to sensor failure, communication loss, or occlusion. We address the task of \emph{trajectory inpainting}: reconstructing contiguous missing segments from observed context. We propose a Temporal Convolutional Network (TCN) with symmetric dilation that relaxes the standard causality constraint, allowing each time step to draw on both past and future observations, a property that is essential for inpainting, but absent from forecasting-oriented architectures. The model is trained with a composite loss that combines weighted mean squared error, boundary--continuity penalties, and a smoothness regularizer. Trained on a synthetic dataset of $1,000$ (train), $200$ (validation), and $300$ (test) two-dimensional trajectories with randomly placed 20% masked segments, the model achieves good R$^{2}$, MSE and MAE metrics.

Problem

• trajectory inpainting
• missing data
• trajectory reconstruction
• incomplete trajectories

Innovation

• Temporal Convolutional Network
• trajectory inpainting
• symmetric dilation
• composite loss
• missing data reconstruction