Companion Mobility to Assist in Future Human Location Prediction

Ngo, Quan T.; Lan, Doi Thi; Yoon, Seokhoon; Jung, Woo‐Sik; Yoon, Taehyun; Yoo, Dae-Seung

doi:10.1109/access.2022.3186319

Cited by 4 publications

(2 citation statements)

References 41 publications

(45 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Location-acquisition tools like the GPS, smartphone, and sensors have helped generate massive quantities of position data from moving objects. The diversity in trajectory data encourages a lot of research on trajectory data mining, such as human trajectory prediction [1]- [3], user route recommendation [4]- [7], and trajectory clustering [8]- [10]. In recent years, anomalous trajectory detection has become an important research topic in many applications.…”

Section: Introductionmentioning

confidence: 99%

Anomalous Indoor Human Trajectory Detection Based on the Transformer Encoder and Self-Organizing Map

Lan,

Yoon

2023

IEEE Access

Self Cite

View full text Add to dashboard Cite

Anomalous human trajectory detection is a critical task in security surveillance in working areas. To identify anomalous human trajectories, understanding features of their movement plays an important role. Therefore, in this work, a Transformer encoder and self-organizing map-based model called TENSO is proposed to learn trajectory characteristics for detecting anomalies. In particular, the proposed model learns the internal characteristics of normal trajectories and clusters of normal trajectory representations in a latent space. To learn the internal characteristics of normal trajectories, the encoder of Transformer with a self-attention mechanism first encodes trajectories into sequences of embedding vectors of trajectory points in the latent space. Then, a decoder reconstructs the trajectories from the latent space. In addition, to learn clusters of normal trajectory representations in the latent space, the self-organizing map (SOM) layer is used, which gets its input as the output of the Transformer encoder. In the training phase, the TENSO model is trained using a total loss of trajectory reconstruction and SOM losses. In the anomaly detection phase, a test trajectory is evaluated to determine whether it is an anomaly based on trajectory reconstruction errors and the quantization error on the SOM. In this phase, a new metric is proposed which, namely WS, is the weighted sum of recall and precision to choose the appropriate threshold for detecting anomalies. The TENSO modelbased framework is evaluated using two real trajectory datasets: MIT Badge and sCREEN. Experimental results show that the proposed framework identifies anomalies effectively and outperforms the baselines.INDEX TERMS Anomalous trajectory detection, indoor human trajectory, Transformer encoder, Selforganizing map.

show abstract

Section: Introductionmentioning

confidence: 99%

Anomalous Indoor Human Trajectory Detection Based on the Transformer Encoder and Self-Organizing Map

Lan,

Yoon

2023

IEEE Access

Self Cite

View full text Add to dashboard Cite

show abstract

“…• In the first phase, in contrast to Icrowd's reliance on the assumption of worker movement following an inhomogeneous Poisson process, we introduce a more advanced approach by employing a recurrent neural net-work (RNN)-based model for worker mobility prediction. RNNs have proven to be highly effective in handling time series data and capturing the intricate patterns of worker movement [18]- [20]. By leveraging the power of RNNs, CAMP overcomes the shortcomings of the inhomogeneous Poisson process assumption, enabling the framework to make predictions more accurately and select workers with a higher probability of visiting locations with a substantial number of tasks.…”

Section: Introductionmentioning

confidence: 99%

Context-Aware Worker Recruitment for Mobile Crowd Sensing Based on Mobility Prediction

Ngo,

Yoon

2023

IEEE Access

Self Cite

View full text Add to dashboard Cite

Opportunistic worker (OW) selection is a challenging problem in mobile crowd sensing (MCS), where tasks are assigned to individuals to be completed seamlessly during their daily routines without any deviation from their usual routes. In this paper, we propose a novel framework named context-aware worker recruitment based on a mobility prediction model (CAMP) to address the OW selection problem in MCS. Unlike previous approaches that relied on worker mobility prediction models with limited accuracy or utility-based selection methods neglecting task distribution differences across locations, CAMP introduces a two-phase strategy for OW selection. In the first phase, we leverage a recurrent neural network-based prediction model specifically designed to forecast volunteer workers' future locations with higher precision. This enhanced mobility prediction ensures more effective task assignments in the MCS system. In the second phase, CAMP employs a weighted-utility algorithm that takes into account the varying task distribution throughout the day across different locations. The key novelty of the CAMP framework lies in its combination of an accurate multi-output RNN model for predicting worker mobility and a unique weighted-utility worker selection algorithm that considers variations in task distribution across different locations and sensing cycles. To validate the effectiveness of the CAMP framework, we extensively evaluate it using real-world GPS data, specifically the Crawdad Roma/Taxi dataset. The results demonstrate that CAMP outperforms existing approaches, delivering a higher number of completed tasks while adhering to the same budget constraints.INDEX TERMS Human mobility prediction, opportunistic worker selection, mobile crowd sensing, task allocation.

show abstract