KLKehan Li

Kehan Li

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Kehan Li has a total of 33 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • High-Resolution Large-Scale Urban Traffic Speed Estimation With Multi-Source Crowd Sensing Data

    by Yingqian Zhang, Chao Li, Kehan Li, Shibo He, Jiming Chen

    High-resolution large-scale urban traffic speed estimation is vital for intelligent traffic management and urban planning. However, single-source data from commonly used sources like cameras, loop detectors, or onboard devices exhibit limitations due to uneven distribution and significant noise, especially in large-scale urban areas. Consequently, existing approaches relying on these single-source data often yield low-resolution and biased estimations. In this study, we take the first attempt to leverage mobile pedestrian data and car navigation data for multi-source fusion, proposing a model to achieve high-resolution urban traffic speed estimation in large-scale areas. The key questions are how to obtain and utilize relatively static roadside pedestrian crowd sensing data to characterize the speed of moving vehicles, and how to design multi-source heterogeneous data fusion framework to improve the overall estimation performance. Specifically, a meta-learning-based matrix decomposition algorithm is first proposed to impute the missing values adaptively considering history speed data. After obtaining the imputed data, we utilize the self-view speed aggregation algorithm learning from complete spatial information to correct the imputed values. Subsequently, a multi-view speed aggregation algorithm is proposed to fuse multi-source data for tracking actual road conditions which improves road coverage. We evaluated our model with real-world datasets collected from more than 500,000 smartphones in Wenzhou, China. Experimental results show that the proposed model outperforms the state-of-the-art approaches by 7.48% and 6.99% in MAPE on missing data imputation and multi-source data fusion models, respectively.


  • Open Access

    Use of temporal contact graphs to understand the evolution of COVID-19 through contact tracing data

    by Mincheng Wu, Chao Li, Zhangchong Shen, Shibo He, Lingling Tang, Jie Zheng, Yi Fang, Kehan Li, Yanggang Cheng, Zhiguo Shi, et al.

    Digital contact tracing has been recently advocated by China and many countries as part of digital prevention measures on COVID-19. Controversies have been raised about their effectiveness in practice as it remains open how they can be fully utilized to control COVID-19. In this article, we show that an abundance of information can be extracted from digital contact tracing for COVID-19 prevention and control. Specifically, we construct a temporal contact graph that quantifies the daily contacts between infectious and susceptible individuals by exploiting a large volume of location-related data contributed by 10,527,737 smartphone users in Wuhan, China. The temporal contact graph reveals five time-varying indicators can accurately capture actual contact trends at population level, demonstrating that travel restrictions (e.g., city lockdown) in Wuhan played an important role in containing COVID-19. We reveal a strong correlation between the contacts level and the epidemic size, and estimate several significant epidemiological parameters (e.g., serial interval). We also show that user participation rate exerts higher influence on situation evaluation than user upload rate does, indicating a sub-sampled dataset would be as good at prediction. At individual level, however, the temporal contact graph plays a limited role, since the behavior distinction between the infected and uninfected individuals are not substantial. The revealed results can tell the effectiveness of digital contact tracing against COVID-19, providing guidelines for governments to implement interventions using information technology.


  • Interblock Flow Prediction With Relation Graph Network for Cold Start on Bike-Sharing System

    by Mingda Jiang, Chao Li, Kehan Li, Zidong Yang, Hao Liu

    As the IoT technology becomes well established and sharing economy expands worldwide, the bike-sharing system (BSS) has spread fast in the last decade. When introducing the BSS in a new city, the operator often faces many challenges: e.g., optimizing station siting (physical or electric), constructing bike lanes, and making strategies for the initial distribution and rebalancing of bikes. These challenges require an accurate interblock flow prediction before deploying the BSS. In this article, we derive blocks from the urban road network and extract blocks’ features based on the distribution and types of POI. Then, the interblock flow can be predicted based on the features of both start/end block and neighbor blocks. We propose a unified architecture with a generalized attention mechanism named the relation graph network (RGN) to extract block features and make predictions across cities. The evaluations on real-world data sets show that RGN outperforms other popular graph neural networks in multiple metrics. We further simulate the applications of station site selection and bike lane planning based on the prediction from RGN, and the result provides a noticeable increase in the BSS’s deployment/operation efficiency.


  • Boost Spectrum Prediction With Temporal-Frequency Fusion Network via Transfer Learning

    by Kehan Li, Chao Li, Jiming Chen, Qiming Zhang, Zebo Liu, Shibo He

    Modeling and predicting the radio spectrum is vital for spectrum management, such as spectrum sharing and anomaly detection. Nevertheless, the precise spectrum prediction is challenging due to the interference from both intra-spectrum and external factors. To tackle these complex internal and external correlations, we develop a model named TF $^2$ AN, consisting of three components: 1) a robust signal detection algorithm based on image processing, 2) an attention-based Long Short-term Memory network to capture the temporal-frequency correlations, 3) a generalized fusion module to take the heterogeneous external factors into account. This structure shows prominent effectiveness for spectrum prediction on a single monitoring station with sufficient data. However, when the data derived from a single station is insufficient, the performance of the deep learning model will decline a lot. Considering that more than one monitoring station is deployed in practice, the new challenge becomes how to enhance our model by leveraging the data from multiple stations or frequency bands. Therefore, we further propose T-TF $^2$ AN, a transfer learning-based framework for data augmentation and knowledge sharing in spectrum prediction. Compared to TF $^2$ AN, better performance is achieved. Besides, the model interpretability and training efficiency are also discussed with two case studies, respectively.


  • Model and Transfer Spatial-Temporal Knowledge for Fine-Grained Radio Map Reconstruction

    by Kehan Li, Chao Li, Baosheng Yu, Zhangchong Shen, Qiming Zhang, Shibo He, Jiming Chen

    Radio map, serving as an efficient indicator of wireless environments, has been widely used in many applications such as network monitoring, anomaly signal detection, and localization. It is hard to maintain an update-to-date fine-grained radio map within a large area since the radio map changes rapidly due to internal and external factors. In this paper, we develop a fine-grained radio map reconstruction framework, called Supreme, based on crowdsourced data in an image super-resolution manner. Specifically, Supreme explores spatial-temporal relationships in historical coarse-grained radio maps and builds a real-time fine-grained radio map via a spatial-temporal reconstruction network. Furthermore, a heterogeneous data fusion module is devised to handle external information. Supreme has shown its superiority on real-world datasets. However, it is costly to collect enough fine-grained radio maps as labels considering the labor-intensive collection process and privacy issues. Therefore, we further present the Supreme-UDA that adapts the model from a labeled source domain towards an unlabeled target domain with the help of a domain classifier. Compared with Supreme, the enhanced model effectively transfers spatial-temporal knowledge and improves generalization ability in new scenes even without any annotations. Finally, a case study illustrates that our model can boost localization accuracy in practice.


  • Destination Prediction Based on Virtual POI Docks in Dockless Bike-Sharing System

    by Mingda Jiang, Chao Li, Kehan Li, Hao Liu

    As the sharing economy develops and bike-sharing apps emerge, the dockless bike-sharing system (DLBS) has become a competitive alternative to the docked bike-sharing system because of its convenience of finding and parking without physical docks. Meanwhile, new demands are rapidly increasing as DLBS expands, e.g., crowd-sourced re-balancing and pre-ordering during rush hours. A more fine-grained destination prediction is required to tackle these issues. In this paper, we propose a probabilistic-trip-based destination prediction method named P 3 M. To overcome the uncertainty due to docks’ absence, we introduce the virtual docks derived from POIs and convert a single trip recorded in GPS into several probabilistic trips among POIs using an innovative user behavior model Walking-Riding-Walking Probabilistic Trip. To deal with sparsity, P 3 M adapts a trip-wised parameter share strategy together with a statistical-based history-feature extractor for better performance without overfitting. Compared with the baseline method, P 3 M reduces the mean absolute errors measured with distance by 31.55% (from 1.1036 km to 0.7554 km) and is less sensitive to the sparsity of user’s records. Further, we analyze the application of P 3 M in different types of DLBS and use two simulations to prove its efficiency under insufficient bike supply circumstances.


  • Supreme: Fine-grained Radio Map Reconstruction via Spatial-Temporal Fusion Network

    by Kehan Li, Jiming Chen, Baosheng Yu, Zhangchong Shen, Chao Li, Shibo He

    Radio map, serving as an efficient indicator of wireless environments, has been widely used in smart-city applications, including network monitoring/planning, anomaly signal detection, and indoor/outdoor localization. It is hard to maintain an update-to-date fine-grained radio map within a large area, since the radio map changes rapidly due to the internal and external factors. Previous studies usually relied on time-consuming site surveys at densely predefined reference points, leading to either coarse-grained or out-of-date radio maps. In this paper, we propose a fine-grained radio map reconstruction framework, called Supreme, based on crowd-sourced data in an image super-resolution manner. Specifically, Supreme explores spatial-temporal relationships in historical coarse-grained radio maps and builds a real-time fine-grained radio map using deep spatial-temporal reconstruction networks. Furthermore, a heterogeneous data fusion module is devised to make full use of external information. To evaluate the performance of Supreme, we conduct extensive experiments and ablation studies on a large-scale dataset with a total of six-month data collected from two university campuses. Besides, we investigate the transferability of Supreme in different locations and service networks, showing that the fine-tuned model can largely reduce the training time and achieve better performance. Experimental results demonstrate that our model outperforms state-of-the-art baselines and a case study on the localization is enhanced with marginal improvements on accuracy.