Trip2Vec: a deep embedding approach for clustering and profiling taxi trip purposes

Human activity recognition (HAR) is a broad research area. While there exist solutions based on sensors and vision-based technologies, these solutions suffer from considerable limitations. Thus in order to mitigate or avoid these limitations, device free solutions based on radio signals like (home) WiFi, in particular 802.11 are considered. Recently, channel state information (CSI), available in WiFi 802.11n networks have been proposed for fine-grained analysis. We are able to detect human activities like Walk, Sit, Stand, Run (in the sequel, any human activity used for classification is capitalised, i.e. is denoted by its corresponding label. For example, "standing" is denoted as Stand, the activity "sitting" is denoted by Sit and so on), etc. in a line-of-sight (LOS) scenario and a non-line-of-sight (N-LOS) scenario within an indoor environment. We propose two algorithms-one using a support vector machine (SVM) for classification and another one using a long shortterm memory (LSTM) recurrent neural network. While the former uses sophisticated pre-processing and feature extraction techniques based on wavelet analysis, the latter processes the raw data directly (after denoising). We show that it is possible to characterize activities and/or human body presence with high accuracy and we compare both approaches with regard to accuracy and performance. Furthermore, we extend the experimental setup to detect human falls, too which is a relevant use-case in the context of ambient assisted living (AAL) and show that with the developed algorithms it is possible to detect falls with high accuracy. In addition, we also show that the algorithms can be used to count the number of people in a room based on the CSI-data, which is a first step towards detecting more complex social behavior and activities. Our paper is an extended version of the paper (Damodaran and Schäfer, Device free human activity recognition using wifi channel state

show abstract

“…For a recent example of using contextual information for activity detection, see e.g. (Chen et al 2019).…”

Section: Prior Workmentioning

confidence: 99%

Device free human activity and fall recognition using WiFi channel state information (CSI)

Damodaran

Haruni

Kokhkharova

et al. 2020

CCF Trans. Pervasive Comp. Interact.

View full text Add to dashboard Cite

show abstract

“…Sutskever et al 12 first proposed the method of sequence-to-sequence auto-encoder for machine translation. Chen et al 10 used auto-encoder fusing a variety of different features to get more semantical and discriminative context representation in the latent space, so as to analyze the taxi destination. Liu et al 13 used embedding method for real-time personalized search and similar product list recommendation.…”

Section: Embeddingmentioning

confidence: 99%

“…The typical problem is that the potential feature information and the association between different types of combined features cannot be found. 22 Therefore, this article uses auto-encoder 10,28 to solve this problem. Auto-encoder belongs to the embedding method, which is an unsupervised artificial neural network.…”

Section: Feature Extraction and Transformationmentioning

confidence: 99%

See 1 more Smart Citation

Unsupervised learning trajectory anomaly detection algorithm based on deep representation

Wang

Yuan

Pei

et al. 2020

International Journal of Distributed Sensor Networks

View full text Add to dashboard Cite

Without ground-truth data, trajectory anomaly detection is a hard work and the result lacks of interpretability. Moreover, in most current methods, trajectories are represented by geometric features or their low-dimensional linear combination, and some hidden features and high-dimensional combined features cannot be found efficiently. Meanwhile, traditional methods still cannot get rid of the limitation of space attributes. Therefore, a novel trajectory anomaly detection algorithm is present in this article. Unsupervised learning mechanism is used to overcome nonground-truth problem and deep representation method is used to represent trajectories in a comprehensive way. First, each trajectory is partitioned into segments according to its open angles, then the shallow features at each point of a segment are extracted and. In this way, each segment is represented as a feature sequence. Second, shallow features are integrated into auto-encoder-based deep feature fusion model, and the fusion feature sequences can be extracted. Third, these fused feature sequences are grouped into different clusters using a unsupervised clustering algorithm, and then segments which quite differ from others are detected as anomalies. Finally, comprehensive experiments are conducted on both synthetic and real data sets, which demonstrate the efficiency of our work.

show abstract

“…On the other hand, these patterns can also reveal travel purposes by supplementing other information. Some research has been conducted to recognize trip purposes from taxi trajectories, and main trip purposes were distinguished, including work-related, transportation transfer, dining, shopping, recreation, and schooling [20,[47][48][49]. Each pattern of departures, arrivals, and spatial interactions discovered by STC-NMF was also related to some specific trip purposes.…”

Section: Spatial Interaction Patternsmentioning

confidence: 99%

A Spatiotemporal Constraint Non-Negative Matrix Factorization Model to Discover Intra-Urban Mobility Patterns from Taxi Trips

Gao

Liu

et al. 2019

Sustainability

View full text Add to dashboard Cite

Taxi services provide an urban transport option to citizens. Massive taxi trajectories contain rich information for understanding human travel activities, which are essential to sustainable urban mobility and transportation. The origin and destination (O-D) pairs of urban taxi trips can reveal the spatiotemporal patterns of human mobility and then offer fundamental information to interpret and reform formal, functional, and perceptual regions of cities. Matrices are one of the most effective models to represent taxi trajectories and O-D trips. Among matrix representations, non-negative matrix factorization (NMF) gives meaningful interpretations of complex latent relationships. However, the independence assumption for observations is violated by spatial and temporal autocorrelation in taxi flows, which is not compensated in classical NMF models. In order to discover human intra-urban mobility patterns, a novel spatiotemporal constraint NMF (STC-NMF) model that explicitly solves spatial and temporal dependencies is proposed in this paper. It factorizes taxi flow matrices in both spatial and temporal aspects, thus revealing inherent spatiotemporal patterns. With three-month taxi trajectories harvested in Beijing, China, the STC-NMF model is employed to investigate taxi travel patterns and their spatial interaction modes. As the results, four departure patterns, three arrival patterns, and eight spatial interaction patterns during weekdays and weekends are discovered. Moreover, it is found that intensive movements within certain time windows are significantly related to region functionalities and the spatial interaction flows exhibit an obvious distance decay tendency. The outcome of the proposed model is more consistent with the inherent spatiotemporal characteristics of human intra-urban movements. The knowledge gained in this research would be useful to taxi services and transportation management for promoting sustainable urban development.Sustainability 2019, 11, 4214 2 of 22 knowledge [6], including human mobility patterns [7,8], driving strategies of cabdrivers for seeking passengers [9][10][11][12], and urban spatial structures [13][14][15], for understanding human mobility and even improving sustainable urban transport.A taxi trip depends on the specific travel purpose. The origin and destination (O-D) points of a trip are considered the critical elements to represent mobility motions, while the route itself is less semantically representative for passengers. Generally, an O-D pair is exactly the pair of pick-up and drop-off points of a taxi ride, which represents the starting and stopping locations of a complete individual trip trajectory. O-D flows generalize not only trajectory geometries but also mobility semantics. Thus, previous studies have explored them to understand human mobility and urban structures. Direction and distance distributions of taxi trips were investigated to model the geographical heterogeneity and distance decay effect of intra-urban human mobility [8]. Mobility hotspots [16], movemen...

show abstract

Trip2Vec: a deep embedding approach for clustering and profiling taxi trip purposes

Cited by 48 publications

References 41 publications

Device free human activity and fall recognition using WiFi channel state information (CSI)

Device free human activity and fall recognition using WiFi channel state information (CSI)

Unsupervised learning trajectory anomaly detection algorithm based on deep representation

A Spatiotemporal Constraint Non-Negative Matrix Factorization Model to Discover Intra-Urban Mobility Patterns from Taxi Trips

Contact Info

Product

Resources

About