Transformer-Based Wildlife Species Classification from Daily Movement Trajectories

2026-05-07

AI Summary

This study addresses the challenge of accurately identifying wildlife species using only everyday movement trajectory data. To this end, we introduce the Transformer architecture—applied here for the first time to wildlife trajectory classification—and model sequences incorporating multidimensional motion features such as displacement, speed, heading, and turning behavior. Experiments are conducted under the original telemetry study’s rigorous evaluation protocol to ensure ecological validity. Results demonstrate that our approach significantly outperforms conventional temporal models—including LSTM, CNN, and temporal convolutional networks—across multiple species and settings. On a binary elephant classification task, the method achieves a balanced accuracy of 0.83 and an AUC of 0.92, representing an 8–22 percentage point improvement over baselines. Performance is optimal with a uniform one-hour temporal resolution, notably enhancing identification accuracy for rare species.

Abstract

Inferring the identity of wildlife species from daily movement data alone is a challenging task. We train sequence models on large-scale, 7-species GPS trajectories from the Movebank platform. Trajectories models are evaluated using a protocol in which entire telemetry studies or regions are heldout during testing. We compare Transformer-based sequence models to LSTM, CNN, and Temporal Convolutional Networks, and find that Transformers consistently achieve higher balanced accuracy with gains of approximately 8 to 22 percentage points, depending on the species and experimental setting. In an elephant binary classification task with 1-hour resolution, the Transformer achieves a balanced accuracy of 0.83 and an AUC of 0.92, substantially outperforming all baseline models. We examine, under data-limited conditions, feature representations by analyzing the differences between a basic displacement-based encoding and an expanded range of movement descriptors that include speed, direction, and turning behavior. With feature augmentation, we see clear performance gains, especially for underrepresented and sparsely represented species, such as large carnivores, lions, and Zebras. Finally, experiments comparing 1-hour and 30-minutetemporal resolutions show that while finer sampling can capture short-term movement patterns for some species, a unified 1-hour resolution yields more promising performance across studies by reducing missing data and ensuring consistent temporal coverage.

Problem

• wildlife species classification
• movement trajectories
• GPS telemetry
• species identification
• animal movement

Innovation

• Transformer
• wildlife species classification
• movement trajectories
• feature augmentation
• temporal resolution