Can we reconstruct accurate travel trajectories from sparse and noisy GPS data?

To address this challenge, researchers at Korea Advanced Institute of Science and Technology (KAIST) and Donghai Laboratory developed a new model called ProChunkFormer, which reconstructs vehicle trajectories from sparse and noisy GPS data, enabling more accurate mobility analysis and intelligent transportation planning.     

They published their study on 2 August 2025, in Communications in Transportation Research.

 

When GPS data is collected at long intervals or suffers from signal dropouts, large gaps emerge in the recorded vehicle trajectory. These gaps make it challenging to reconstruct the true path of travel, especially in dense urban road networks with many alternatives. ProChunkFormer was designed to address this problem by breaking down the reconstruction task into smaller, more manageable steps — progressively refining the predicted trajectory from a coarse outline to detailed local paths.

 

Progressive decoding improves long-interval trajectory reconstruction

In the study, the research team designed a two-stage process to recover complete travel trajectories from GPS data sampled as infrequently as once every four minutes. First, a “skeleton trajectory” is generated to provide a coarse outline of the trip. Then, missing segments between anchor points are reconstructed in parallel using a fine-grained decoder. Unlike prior models that suffer from cumulative errors and require full sequence processing, ProChunkFormer handles long trajectories efficiently by decomposing them into manageable chunks.

 

“Traditional models either interpolate sparsely or struggle with long-distance reconstruction, especially when signal loss occurs in tunnels or high-rise areas,” says Dr. Yonghui Liu, a researcher at KAIST and first author of the paper. “Our approach mimics how humans might infer a trip from key locations — first outlining the major path, then filling in the details — and this progressive structure proves highly effective.”

 

Trajectory gaps bridged even under severe sparsity

The study also revealed how ProChunkFormer handles real-world data sparsity with improved robustness. For example, when tested on GPS traces with sampling intervals of up to 240 seconds — a level at which most methods fail — the model continued to generate accurate and continuous trajectories aligned with actual road networks. This capability is particularly valuable for data collected from battery-limited sensors or privacy-constrained services, where data completeness cannot be guaranteed.

 

“Many public datasets or commercial services only offer limited trajectory points due to privacy or power-saving needs,” explains Li Qian, a researcher at Donghai Laboratory and co-author of the study. “With ProChunkFormer, even a handful of points can be turned into a coherent, road-consistent trajectory.”

 

From sparse data to smart decisions

Accurately reconstructed trajectories open new doors in urban mobility analytics. ProChunkFormer enables more reliable travel time estimation, traffic demand prediction, and even mode detection from incomplete GPS logs. This enhances the utility of existing mobility datasets and reduces the reliance on expensive high-frequency tracking.

 

The authors suggest that the model has the potential to support real-time applications in navigation and traffic control systems by recovering lost paths on the fly, offering resilience to GPS dropout in tunnels, urban canyons, or data-poor regions.

 

The above research is published in Communications in Transportation Research (COMMTR), which is a fully open access journal co-published by Tsinghua University Press and Elsevier. COMMTR publishes peer-reviewed high-quality research representing important advances of significance to emerging transport systems. COMMTR is also among the first transportation journals to make the Replication Package mandatory to facilitate researchers, practitioners, and the general public in understanding and advancing existing knowledge. At its discretion, Tsinghua University Press will pay the open access fee for all published papers in 2025.

 

About Communications in Transportation Research

Communications in Transportation Research was launched in 2021, with academic support provided by Tsinghua University and China Intelligent Transportation Systems Association. The Editors-in-Chief are Professor Xiaobo Qu, a member of the Academia Europaea from Tsinghua University and Professor Shuai’an Wang from Hong Kong Polytechnic University. The journal mainly publishes high-quality, original research and review articles that are of significant importance to emerging transportation systems, aiming to serve as an international platform for showcasing and exchanging innovative achievements in transportation and related fields, fostering academic exchange and development between China and the global community.

It has been indexed in SCIE, SSCI, Ei Compendex, Scopus, CSTPCD, CSCD, OAJ, DOAJ, TRID and other databases. It was selected as Q1 Top Journal in the Engineering and Technology category of the Chinese Academy of Sciences (CAS) Journal Ranking List. In 2022, it was selected as a High-Starting-Point new journal project of the “China Science and Technology Journal Excellence Action Plan”. In 2024, it was selected as the Support the Development Project of “High-Level International Scientific and Technological Journals”. The same year, it was also chosen as an English Journal Tier Project of the “China Science and Technology Journal Excellence Action Plan Phase Ⅱ”. In 2024, it received the first impact factor (2023 IF) of 12.5, ranking Top1 (1/58, Q1) among all journals in “TRANSPORTATION” category. In 2025, its 2024 IF was announced as 14.5, maintaining the Top 1 position (1/61, Q1) in the same category. Tsinghua University Press will cover the open access fee for all published papers in 2025.

---

---