A traffic pattern detection algorithm based on multimodal sensing

Abstract

Nowadays, smartphones are widely and frequently used in people’s daily lives for their powerful functions, which generate an enormous amount of data accordingly. The large volume and various types of data make it possible to accurately identify people’s travel behaviors, that is, transportation mode detection. Using the transportation mode detection, results can increase commuting efficiency and optimize metropolitan transportation planning. Although much work has been done on transportation mode detection problem, the accuracy is not sufficient. In this article, an accurate traffic pattern detection algorithm based on multimodal sensing is proposed. This algorithm first extracts various sensory features and semantic features from four types of sensor (i.e. accelerator, gyroscope, magnetometer, and barometer). These sensors are commonly embedded in commodity smartphones. All the extracted features are then fed into a convolutional neural network to infer traffic patterns. Extensive experimental results show that the proposed scheme can identify four transportation patterns with 94.18% accuracy.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 13 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.