An Enhanced Transportation Mode Detection Method Based on GPS Data

Jing, Liang; Zhu, Qiuhui; Zhu, Min; Li, Mingzhao; Li, Xiaowei; Wang, Jianhua; You, Silan; Zhang, Yilan

doi:10.1007/978-981-10-6385-5_51

Cited by 10 publications

(17 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The performance of these studies has produced significant results, but it has a big limitation that it is not easy to carry multiple sensors all the time. Few studies [20, 29, 35, 36] make use of the GIS data with GPS sensor data, but the GIS information is not always available for all places [14], and the structure of city changes from time to time. This results in the change of GIS maps and in result need to re‐train the models every time the update is recorded in the GIS data.…”

Section: Related Workmentioning

confidence: 99%

Convolutional LSTM based transportation mode learning from raw GPS trajectories

Nawaz

Huang

Wang

et al. 2020

IET Intelligent Transport Systems

View full text Add to dashboard Cite

Section: Related Workmentioning

confidence: 99%

Convolutional LSTM based transportation mode learning from raw GPS trajectories

Nawaz

Huang

Wang

et al. 2020

IET Intelligent Transport Systems

View full text Add to dashboard Cite

“…Finally, non-walk episodes are further classified into travel by car, bus or other modes based on probability model or machine learning method. Meanwhile, an enhanced transition point identification method [ 83 ] was proposed on the basis of the work in Ref. [ 79 ] to identify the transportation mode of segments by using Random Forest-based detection model.…”

Section: Gps Data Classification Based On the Transportation Modementioning

confidence: 99%

“…[ 84 ], authors in Ref. [ 83 ] first merge segments whose distance meets the threshold merely into its following segment instead of preceding and following segments to identify more transition points. Then, adopting iterative method in the segment collection to improve the accuracy of detecting transition point.…”

Section: Gps Data Classification Based On the Transportation Modementioning

confidence: 99%

A Review of GPS Trajectories Classification Based on Transportation Mode

Yang

Stewart

Tang

et al. 2018

Sensors

View full text Add to dashboard Cite

GPS trajectories generated by moving objects provide researchers with an excellent resource for revealing patterns of human activities. Relevant research based on GPS trajectories includes the fields of location-based services, transportation science, and urban studies among others. Research relating to how to obtain GPS data (e.g., GPS data acquisition, GPS data processing) is receiving significant attention because of the availability of GPS data collecting platforms. One such problem is the GPS data classification based on transportation mode. The challenge of classifying trajectories by transportation mode has approached detecting different modes of movement through the application of several strategies. From a GPS data acquisition point of view, this paper macroscopically classifies the transportation mode of GPS data into single-mode and mixed-mode. That means GPS trajectories collected based on one type of transportation mode are regarded as single-mode data; otherwise it is considered as mixed-mode data. The one big difference of classification strategy between single-mode and mixed-mode GPS data is whether we need to recognize the transition points or activity episodes first. Based on this, we systematically review existing classification methods for single-mode and mixed-mode GPS data and introduce the contributions of these methods as well as discuss their unresolved issues to provide directions for future studies in this field. Based on this review and the transportation application at hand, researchers can select the most appropriate method and endeavor to improve them.

show abstract

“…Based on the success of deep learning techniques in recent era, we choose deep net based architecture to solve our problem of GPS trajectory completion only based on GPS data. A few studies, like References [ 14 , 27 ], use data from other sensors like accelerometer, gyroscope, WiFi with the GPS data, and similarly few studies makes use of GIS information with the GPS data [ 28 ] for trajectory mining tasks. It is not always feasible to carry multiple sensors, or WiFi is not necessarily available at all locations.…”

Section: Introductionmentioning

confidence: 99%

GPS Trajectory Completion Using End-to-End Bidirectional Convolutional Recurrent Encoder-Decoder Architecture with Attention Mechanism

Nawaz

Huang

Wang

et al. 2020

Sensors

View full text Add to dashboard Cite

GPS datasets in the big data regime provide rich contextual information that enable efficient implementation of advanced features such as navigation, tracking, and security in urban computing systems. Understanding the hidden patterns in large amount of GPS data is critically important in ubiquitous computing. The quality of GPS data is the fundamental key problem to produce high quality results. In real world applications, certain GPS trajectories are sparse and incomplete; this increases the complexity of inference algorithms. Few of existing studies have tried to address this problem using complicated algorithms that are based on conventional heuristics; this requires extensive domain knowledge of underlying applications. Our contribution in this paper are two-fold. First, we proposed deep learning based bidirectional convolutional recurrent encoder-decoder architecture to generate the missing points of GPS trajectories over occupancy grid-map. Second, we interfaced attention mechanism between enconder and decoder, that further enhance the performance of our model. We have performed the experiments on widely used Microsoft geolife trajectory dataset, and perform the experiments over multiple level of grid resolutions and multiple lengths of missing GPS segments. Our proposed model achieved better results in terms of average displacement error as compared to the state-of-the-art benchmark methods.

show abstract

An Enhanced Transportation Mode Detection Method Based on GPS Data

Cited by 10 publications

References 13 publications

Convolutional LSTM based transportation mode learning from raw GPS trajectories

Convolutional LSTM based transportation mode learning from raw GPS trajectories

A Review of GPS Trajectories Classification Based on Transportation Mode

GPS Trajectory Completion Using End-to-End Bidirectional Convolutional Recurrent Encoder-Decoder Architecture with Attention Mechanism

Contact Info

Product

Resources

About