Accurate imputation of missing traffic data remains critical for intelligent transportation systems, but existing methods often inadequately balance global patterns and localized correlations. This study proposes a novel two-stage imputation framework that harmonizes tensor decomposition with deep learning mechanisms. In the first stage, low-rank tensor completion (LRTC) model extracts latent low-rank features from incomplete traffic data, establishing global priors. The second stage fuses the priors with raw observations via a fusion module, feeding the enhanced representation into a Spatial-Temporal Feature Enhancement Network (STFEN) with scalable window attention. Directed search for strongly relevant features in sparse spatiotemporal attention by introducing domain prior knowledge effectively improves accuracy and computational efficiency. A comprehensive evaluation of three real-world traffic datasets demonstrates the superiority of the two-stage imputation approach compared to state-of-the-art baselines under different missing scenarios. We further systematically evaluate seven matrix/tensor decomposition variants as potential candidates for the first stage processor, with experimental results demonstrating that superior first stage model accuracy enables greater precision recovery in the subsequent phase. This work establishes a new methodological framework to enhance sparse traffic data imputation, thereby facilitating the deployment of data-driven architectures in urban computing systems.

In the digital era, the increasing demand for network traffic necessitates strategic network infrastructure planning. Accurate modeling of traffic demand through cellular traffic generation is crucial for optimizing base station deployment, enhancing network efficiency, and fostering technological innovation. In this paper, we introduce STOUTER, a spatio-temporal diffusion model for cellular traffic generation. STOUTER incorporates noise into traffic data through a forward diffusion process, followed by a reverse reconstruction process to generate realistic cellular traffic. To effectively capture the spatio-temporal patterns inherent in cellular traffic, we pre-train a temporal graph and a base station graph, and design the Spatio-Temporal Feature Fusion Module (STFFM). Leveraging STFFM, we develop STUnet, which estimates noise levels during the reverse denoising process, successfully simulating the spatio-temporal patterns and uncertainty variations in cellular traffic. Extensive experiments conducted on five cellular traffic datasets across two regions demonstrate that STOUTER improves cellular traffic generation by 52.77% in terms of the Jensen-Shannon Divergence (JSD) metric compared to existing models. These results indicate that STOUTER can generate cellular traffic distributions that closely resemble real-world data, providing valuable support for downstream applications.

With the rapid development of new smart devices, such as smart home appliances and VR/AR equipment, there is an increasing demand for novel interaction methods. However, many existing interaction methods require external devices, are unintuitive, and demand substantial user learning effort. To fill this gap, we propose ArmPad, a system that leverages the smartwatch's built-in IMU to enablemultidimensional inputon the user's forearm. Methodologically, ArmPad is explicitly designed to address three core research challenges in forearm-based interaction. First, to resolve the inherent feature conflicts between discrete gesture recognition and continuous distance estimation, we propose a multi-task learning framework with a dynamic gating mechanism for cross-task synergy. Second, to tackle the physical limitation of rapid vibration attenuation across the forearm, we introduce a cross-device guidance strategy that incorporates high-fidelity fingertip knowledge during the training phase. Finally, to ensure robust generalization across diverse populations, we develop task-specific data augmentation and a lightweight user registration mechanism to effectively mitigate physiological variances. Experiments on 20 subjects demonstrate that ArmPad achieves an accuracy of 92.51% on nine gestures and a Mean Absolute Error of 1.53$cm$for sliding distance in cross-user settings. Extensive robustness evaluations and case studies further confirm the system's stability and usability under diverse real-world conditions.

The accurate estimated time of arrival (ETA) is crucial for mobility and transportation applications. Although significant efforts have been made to improve ETA prediction, most existing approaches ignore the influence of individual driving habits and preferences, known as the driving style. Since different drivers may prefer specific routes and speeds based on their experience and familiarity with traffic conditions, driving styles play a crucial role in determining the actual ETA. To fill this gap, we present a novel approach, DSETA, which leverages deep learning to learn and then integrate driving style representations for personalized and precise ETA predictions. Our method employs a diffusion model that captures nuanced driving styles by generating driving speed distribution. We also utilize attention mechanisms to dynamically adjust the impacts of various spatio-temporal factors and driving styles on ETA predictions. Additionally, we introduce a Multi-View Multi-Task framework that incorporates auxiliary tasks, including segment-view driving style classification and route-view speed distribution prediction, to enhance the ETA learning process. A route-level speed prior regularization strategy further improves the model's generalization capabilities. Extensive experiments conducted on a large real-world trip trajectory dataset demonstrate that DSETA achieves high effectiveness and outperforms various baselines across multiple evaluation metrics.

Accurate modeling of traffic demand through cellular traffic generation is crucial for optimizing base station deployment. We thus present STOUTER, a Spatio-Temporal diffusiOn model for cellUlar Traffic genERation. To effectively capture spatial and temporal dynamics, we pretrain both a temporal graph and a base-station graph, and introduce a Spatio-Temporal Feature Fusion Module (STFFM). On five datasets from two regions, STOUTER reduces Jensen-Shannon Divergence by 52.8% over prior methods, generating distributions that closely match real traffic and aiding downstream planning tasks.

In traffic flow prediction, the efficacy of deep learning models is largely contingent upon the availability of extensive training datasets, presenting a formidable challenge in data-scarce environments. Transfer learning has emerged as a promising strategy to address this challenge by leveraging abundant data from source cities to enhance predictive accuracy in target cities with limited data. Nonetheless, existing methods frequently neglect the distinct characteristics and interrelationships among various regions within cities, leading to predominantly city-level knowledge transfers that underutilize the potential of transferred information. In this paper, we present SERT, a fine-grained regional knowledge transfer method specifically designed to mitigate data scarcity in traffic flow prediction. SERT initiates the process by establishing relationships between source and target regions through the integration of satellite imagery and Points of Interest (POI) data, effectively capturing region-specific features to create matched region pairs. Subsequently, we propose an innovative contrastive domain adaptation strategy to align the features of these matched regions, thereby facilitating inter-regional knowledge transfer while maximizing the feature distance of unmatched regions to reduce interference from irrelevant data. This approach enables the effective transfer of valuable knowledge from the source cities to its relevant counterparts in the target city. Comprehensive experimental results demonstrate that SERT outperforms existing methods in terms of prediction accuracy while ensuring significant computational efficiency. The code is available at https://github.com/MobiXg/SERT

Map matching aims to align GPS trajectories to their actual travel routes on a road network, which is an essential pre-processing task for most of trajectory-based applications. Many map matching approaches utilize Hidden Markov Model (HMM) as their backbones. Typically, HMM treats GPS samples of a trajectory as observations and nearby road segments as hidden states. During map matching, HMM determines candidate states for each observation with a fixed searching range, and computes the most likely travel route using theViterbialgorithm. Although HMM-based approaches can derive high matching accuracy, they still suffer from high computation overheads. By inspecting the HMM process, we find that the computation bottleneck mainly comes from improper candidate sets, which contain many irrelevant candidates and incur unnecessary computations. In this paper, we present$\mathtt {LiMM}$– a learned road network index structure for efficient map matching.$\mathtt {LiMM}$improves existing HMM-based approaches from two aspects. First, we propose a novel learned index for road networks, which considers the characteristics of road data. Second, we devise an adaptive searching range mechanism to dynamically adjust the searching range for GPS samples based on their locations. As a result,$\mathtt {LiMM}$can provide refined candidate sets for GPS samples and thus accelerate the map matching process. Extensive experiments are conducted with three large real-world GPS trajectory datasets. The results demonstrate that$\mathtt {LiMM}$significantly reduces computation overheads by achieving an average speedup of$11.7\times$than baseline methods, merely with a subtle accuracy loss of 1.8%.

Dynamic ridesharing has gained significant attention in recent years. However, existing ridesharing studies often focus on optimizing order dispatching and vehicle repositioning separately, leading to short-sighted decisions and underutilization of the ridesharing potential. In this paper, we propose a novel joint optimization framework called$\mathtt {JODR}$. By coordinating order dispatching and vehicle repositioning,$\mathtt {JODR}$enhances ridesharing efficiency while ensuring high-quality service. The core idea of$\mathtt {JODR}$is to dispatch ride orders with high demand in specific mobility directions to vehicles with sufficient available capacity, effectively balancing future supply and demand in those directions. To achieve this, we introduce a novel mobility value function that can predict the long-term mobility value of matching an order with its travel direction. By considering orders’ directional mobility values, service quality assessments, and available vehicle capacities,$\mathtt {JODR}$formulates the order dispatching as a minimum-cost maximum-flow problem to derive the optimal order-vehicle assignments. Furthermore, the value function helps the intelligent repositioning of idle vehicles. Extensive experiments conducted on a large real-world dataset demonstrate the superiority of$\mathtt {JODR}$over state-of-the-art methods across various performance metrics. These experimental results validate the effectiveness of$\mathtt {JODR}$in improving the ridesharing efficiency and experience.

Mobile devices have increasingly integrated with numerous deep learning-based visual applications, such as object classification and recognition models. While these models perform well in controlled environments, their effectiveness declines in real-world environment due to out-of-distribution (OOD) data not seen during training. Existing methods for detecting OOD data often compromise normal data recognition and require extensive training on unattainable OOD data. To address these issues, we propose$\mathtt {POD}$, a framework designed to enhance mobile visual applications by providing high-precision OOD detection without affecting original model performance. In the offline phase,$\mathtt {POD}$generates OOD detectors from any classification model by analyzing model's neuron responses to various data types. In the online phase, it continuously adjusts decision boundaries by integrating results from both the original model and the detector. Evaluated on two public datasets and one self-collected dataset across various popular classification models,$\mathtt {POD}$significantly improves OOD detection performance while maintaining the accuracy of original models.

Existing domain generalization (DG) methods for cross-person sensor-based activity recognition tasks often struggle to capture both intra- and inter-domain style diversity, leading to significant domain gaps with the target domain. In this study, we explore a novel perspective to tackle this problem, a process conceptualized as domain padding. This proposal aims to enrich the domain diversity by synthesizing intra- and inter-domain style data while maintaining robustness to class labels. We instantiate this concept using a conditional diffusion model and introduce a style-fused sampling strategy to enhance data generation diversity, termed Diversified Intra- and Inter-domain distributions via activity Style-fused Diffusion modeling (DI2SDiff). In contrast to traditional condition-guided sampling, our style-fused sampling strategy allows for the flexible use of one or more random style representations from the same class to guide data synthesis. This feature presents a notable advancement: it allows for the maximum utilization of possible combinations among existing styles to generate a broad spectrum of new style instances. We further extend DI2SDiff into DI2SDiff++ by enhancing the diversity of style guidance. Specifically, DI2SDiff++ integrates a multi-head style conditioner to provide multiple distinct, decomposed substyles and introduces a substyle-fused sampling strategy that allows cross-class substyle fusion for broader guidance. Empirical evaluations on a wide range of datasets demonstrate that our generated data achieves remarkable diversity within the domain space. Both intra- and inter-domain generated data have been proven significant and valuable, enabling DI2SDiff and DI2SDiff++ to surpass state-of-the-art DG methods in various cross-person activity recognition tasks.

Dynamic ridesharing has garnered significant attention in recent years due to its numerous benefits. Existing ridesharing algorithms often employ a “filter-and-refine” frame-work, where a large set of candidate vehicles is initially selected for each ride order, followed by computationally intensive route planning for each candidate. However, this process can lead to significant response delays and limit system efficiency. To address this challenge, we propose an order-vehicle pre-matching recommendation approach (PreMR) that refines the candidate set before route planning. PreMR leverages spatial-temporal intervals and a self-attention mechanism to encode diverse order and vehicle information into uniform and informative representations, enabling it to accurately identify the most suitable vehicles for each order. Extensive experiments using real-world datasets and four representative ridesharing algorithms demonstrate that PreMR significantly reduces order response time (by 46.78% on average) while maintaining high service quality, with a slight trade-off in the order completion rate.

Existing domain generalization (DG) methods for cross-person generalization tasks often face challenges in capturing intra- and inter-domain style diversity, resulting in domain gaps with the target domain. In this study, we explore a novel perspective to tackle this problem, a process conceptualized as domain padding. This proposal aims to enrich the domain diversity by synthesizing intra- and inter-domain style data while maintaining robustness to class labels. We instantiate this concept using a conditional diffusion model and introduce a style-fused sampling strategy to enhance data generation diversity. In contrast to traditional condition-guided sampling, our style-fused sampling strategy allows for the flexible use of one or more random styles to guide data synthesis. This feature presents a notable advancement: it allows for the maximum utilization of possible permutations and combinations among existing styles to generate a broad spectrum of new style instances. Empirical evaluations on a broad range of datasets demonstrate that our generated data achieves remarkable diversity within the domain space. Both intra- and inter-domain generated data have proven to be significant and valuable, contributing to varying degrees of performance enhancements. Notably, our approach outperforms state-of-the-art DG methods in all human activity recognition tasks.

Federated learning (FL) is an emerging distributed and privacy-preserving machine learning framework. However, the performance of traditional FL methods is seriously impaired by the real-world data, which appear to be non-IID. The recent clustered federated learning (CFL) methods eliminate the impact of non-IID data by grouping clients with similar data distribution into the same cluster. Unfortunately, existing CFL methods heavily rely on the pre-setting of the cluster number, failing to achieve adaptive client clustering. We also experimentally observe that imbalanced data largely degrade their correctness of client clustering. In this paper, we present a novel CFL method without manual intervention, named AutoCFL, which can eliminate both effects of non-IID and imbalanced data simultaneously. To deal with imbalanced data, the local training adjustment strategy adaptively adjusts the number of local training epochs for each client. To further improve the clustering correctness and adaptability, the weighted voting-based client clustering strategy automatically groups each client into an appropriate cluster. Extensive experiments are conducted to evaluate the design of AutoCFL with three popular datasets under various data settings. Experimental results demonstrate that AutoCFL outperforms state-of-the-art methods, e.g., on average improving model accuracy by 9.24%, while reducing communication costs by 4.67 in an adaptive manner.

Clustered federated learning (CFL) has proved to be an effective way to alleviate the non-IID (not independently and identically distributed) data challenge, which severely restricts the wider application of federated learning. However, existing approaches either lack adaptability,i.e., they require an additional number of clusters as a guide when clustering, or lack effectiveness in terms of communication. In this paper, we explore the differences in the ability of different layers in a model to represent non-IID data, and propose a hierarchical CFL approach, namedHiCFL, which considers both adaptivity and communication efficiency. The improvement of communication efficiency is due to our proposed novel concept of model stability, which characterizes the variation of model weights during training. Based on model stability,HiCFLcan find the proper time to bi-partition the clusters of mobile devices in a hierarchical manner more quickly. We conduct extensive experiments based on popular datasets with various non-IID data settings. The results show thatHiCFLachieves excellent performance effectiveness and efficiency. Compared to state-of-the-art approaches,HiCFLcan improve the model accuracy by$2.0\% \sim 9.0\%$, while reducing the communication overheads by$27.3\% \sim 80.6\%$.

Global Positioning System (GPS) has benefited many novel applications, e.g., navigation, ride-sharing, and location-based services, in our daily life. Although GPS works well in most places, its performance in urban canyons is well-known poor, due to the signal reflections of non-line-of-sight (NLOS) satellites. Tremendous efforts have been made to mitigate the impacts of NLOS signals, while previous works heavily rely on precise proprietary 3D city models or other third-party resources, which are not easily accessible. In this paper, we presentDeepGPS, a deep learning enhanced GPS positioning system that can correct GPS estimations by only considering some simple contextual information.DeepGPSfuses environmental factors, including building heights and road distribution around GPS's initial position, and satellite statuses to describe the positioning context, and exploits an encoder-decoder network model to implicitly learn the complex relationships between positioning contexts and GPS estimations from massive labeled GPS samples. As a result, the well-trained model can accurately predict the correct position for each erroneous GPS estimation given its positioning context. We further improve the model with a novel constraint mask to filter out invalid candidate locations, and enable continuous localization with a simple mobility model. A prototype system is implemented and experimentally evaluated using a large-scale bus trajectory dataset and real-field GPS measurements. Experimental results demonstrate thatDeepGPSsignificantly enhances GPS performance in urban canyons, e.g., on average effectively correcting 90.1% GPS estimations with accuracy improvement by 64.6%.

Ride-hailing service (RHS) has become an important transportation mode in our daily life. Although many works have been proposed to improve RHS from different aspects, only few works focus on the selections of pick-up locations, where rider and driver meet and start a trip. In this paper, we presentMPLRec, a data-driven pick-up location recommendation system that exploits riders' specific mobility demands,e.g.destination, and historical experiences to meet riders' travel requirements.MPLRecgenerates potential pick-up locations over the road network and characterizes them with rich features that describe a location from the riders' perspective. We also build spatio-temporal indexes to organize potential pick-up locations and historical data for facilitating online recommending. When processing an online recommendation request,MPLRecderives candidate pick-up locations and investigates them with materialized features, which are computed from historical order and trajectory data while considering rider's mobility demands. Based on these features, a novel scoring function is used to derive the best pick-up location for each request. Moreover, we implement an RHS simulator to evaluateMPLRecusing large-scale practical ride-hailing datasets. Extensive experiments and simulations demonstrate the effectiveness and efficiency ofMPLRec, which can complete each request within 0.5 s and largely reduce the ride-hailing costs when compared to baseline methods.

As an essential component, roadside units (RSUs) play an indispensable role in realizing Vehicle-to-Everything (V2X) by seamlessly connecting various intelligent devices and vehicles. To facilitate the construction of V2X, much research has been done in designing effective RSU deployment strategies. However, most of these efforts are largely limited by design utility and deployment scalability. To address the limitations of previous works, this paper proposes a general RSU deployment framework,Greta, which can evaluate candidate deployment sites from different perspectives with rich input data, and satisfy different requirements on optimization metrics. To this end, we model the general RSU deployment problem as a customized reinforcement learning (RL) problem that intelligently explores the deployment environment to find a good deployment strategy. Specifically, we design an effective data profiling network to extract features from multi-modality input data. These extracted features are gradually weighted, fused, and encoded as part of the state representation of the RL model. We further design new reward functions considering various deployment metrics and propose an action space pruning scheme to speed up model training. We implement a prototype system ofGretaand extensively evaluate its performance using real-world data. The results showGretaachieves remarkable performance gains compared to recent RSU deployment methods.

Graphs have been widely used for modeling large-scale data generated from real-world applications, while compact representation of such graphs is beneficial for efficient storage and effective graph analysis. As a promising solution, lossless graph summarization can compactly represent a given graph as a summary graph, which consists of supernodes (i.e., sets of nodes) and superedges (edges between supernodes), and the correction edge sets, which together with summary graph can exactly reconstruct the original graph. Although many research efforts have been devoted to develop graph summarization methods, existing works are still inefficient in terms of computation efficiency and representation compactness. To address their limitations, we propose optGS that includes a set of optimization techniques, including computation-oriented supernode re-dividing, degree-aware approximation metric for selecting the best merge, and redundant computation avoidance, to improve current advances. Extensive experiments on a variety of large graph datasets demonstrate the computation efficiency and compression effectiveness of our optGS, e.g., improving the representation compactness by up to 20.28% and achieving 3.53× speedup in running time than the state-of-the-art methods.

Domain generalization refers to the challenge of training a model from various source domains that can generalize well to unseen target domains. Contrastive learning is a promising solution that aims to learn domain-invariant representations by utilizing rich semantic relations among sample pairs from different domains. One simple approach is to bring positive sample pairs from different domains closer, while pushing negative pairs further apart. However, in this paper, we find that directly applying contrastive-based methods is not effective in domain generalization. To overcome this limitation, we propose to leverage a novel contrastive learning approach that promotes class-discriminative and class-balanced features from source domains. Essentially, clusters of sample representations from the same category are encouraged to cluster, while those from different categories are spread out, thus enhancing the model's generalization capability. Furthermore, most existing contrastive learning methods use batch normalization, which may prevent the model from learning domain-invariant features. Inspired by recent research on universal representations for neural networks, we propose a simple emulation of this mechanism by utilizing batch normalization layers to distinguish visual classes and formulating a way to combine them for domain generalization tasks. Our experiments demonstrate a significant improvement in classification accuracy over state-of-the-art techniques on popular domain generalization benchmarks, including Digits-DG, PACS, Office-Home and DomainNet.

Characterizing human driver's driving behaviors from GPS trajectories is an important yet challenging trajectory mining task. Previous works heavily rely on high-quality GPS data to learn such driving style representations through deep neural networks. However, they have overlooked the driving contexts that greatly govern drivers' driving activities and the data sparsity issue of practical GPS trajectories collected at a low-sampling rate. Besides, existing works omit the cold start problem, where the newly joined drivers usually have insufficient data to learn accurate driving style representations. To address these limitations, we present an adversarial driving style representation learning approach, named$\mathtt {Radar}$. In addition to summarizing statistic features from raw GPS data,$\mathtt {Radar}$also extracts contextual features from three aspects of road condition, geographic semantic, and traffic condition. We exploit the advanced semi-supervised generative adversarial networks to construct our learning model. By jointly considering statistic features and contextual features, the trained model is able to efficiently learn driving style representations from practical GPS trajectory data. Furthermore, we enhance$\mathtt {Radar}$'s representation learning for drivers owning limited training data with some basic data augmentation strategies and a novel auxiliary driver based data augmentation method. Experiments on two benchmark applications,i.e., driver identification and driver number estimation, with a large real-world GPS trajectory dataset demonstrate that$\mathtt {Radar}$can outperform the state-of-the-art approaches by learning more effective and accurate driving style representations.

Quality-of-Service (QoS) based collaborative prediction models are emerging to select appropriate edge cloud services for users. Nevertheless, there are still challenges in the realworld QoS prediction task. First, existing QoS prediction models are mostly transductive, failing to generalize to unseen users and services. Secondly, an accurate prediction model remains unexplored under the extreme sparse data scenario, where only a few interactions are available for collaborative filtering. To address these problems, we propose -Inductive -Subgraph -Pattern -Aware Graph Neural Network (ISPA-GNN), which leverages a novel graph-based collaborative filtering method with a subgraph sampling strategy. We further optimize the embeddings components, replacing the user/service embeddings with compositional context information to enable better generalization to unseen nodes while reducing memory usage. Extensive experiments on a large-scale real-world service QoS dataset demonstrate some decent properties of our model, including high prediction accuracy, memory efficiency, and slight performance degradation even if 25% of users/services are never seen.

Due to the wide availability of taxis in a city and the tremendous benefits of ridesharing, taxi ridesharing becomes promising and attractive in recent years. Existing taxi ridesharing schemes simply match ride requests and taxis based on partial trip information and omit the offline passengers, who will hail a taxi at the roadside without submitting the ride requests to the system. Thus, they are still not efficient and practical. In this article, we consider the mobility-aware taxi ridesharing problem and presentmT-Shareto address these limitations.mT-Sharefully exploits the mobility information of taxis and ride requests to achieve efficient indexing of taxis/requests and better passenger–taxi matching, while still satisfying the constraints on passengers’ deadlines and taxis’ capacities. Specifically,mT-Sharemakes use of both geographical information and travel directions to index taxis and ride requests and supports the shortest path-based routing and probabilistic routing to serve both online and offline ride requests. In addition,mT-Shareproposes a novel payment model to share the ridesharing benefits among the taxi driver and passengers. Extensive evaluations using a large real-world taxi data set demonstrate the efficiency and effectiveness ofmT-Share, which can respond each ride request in milliseconds and be with moderate detour costs and passengers’ waiting time. Compared to state-of-the-art schemes,mT-Sharecan serve 42% and 62% more ride requests in peak and nonpeak hours, respectively. Furthermore,mT-Sharecan save 8.6% taxi fare for passengers and meanwhile increase 7.8% incomes for taxi drivers, when compared with the regular taxi services.

In this article, we propose and study a novel data-driven framework for Targeted Outdoor Advertising Recommendation (TOAR) with a special consideration of user profiles and advertisement topics. Given an advertisement query and a set of outdoor billboards with different spatial locations and rental prices, our goal is to find a subset of billboards, such that the total targeted influence is maximum under a limited budget constraint. To achieve this goal, we are facing two challenges: (1) it is difficult to estimate targeted advertising influence in physical world; (2) due to NP hardness, many common search techniques fail to provide a satisfied solution with an acceptable time, especially for large-scale problem settings. Taking into account the exposure strength, advertisement matching degree, and advertising repetition effect, we first build a targeted influence model that can characterize that the advertising influence spreads along with users mobility. Subsequently, based on a divide-and-conquer strategy, we develop two effective approaches, i.e., a master–slave-based sequential optimization method, TOAR-MSS, and a cooperative co-evolution-based optimization method, TOAR-CC, to solve our studied problem. Extensive experiments on two real-world datasets clearly validate the effectiveness and efficiency of our proposed approaches.

Proactive taxi dispatching is of great importance to balance taxi demand-supply gaps among different locations in a city. Recent advances primarily rely on deep reinforcement learning (DRL) to directly learn the optimal dispatching policy. These works, however, are still not sufficiently efficient because they overlook several pieces of valuable context information. As a result, they may generate quite a few improper actions and introduce unnecessary coordination costs. To improve existing works, we presentCOX– a context-aware taxi dispatching approach that incorporates rich contexts into DRL modeling for more efficient taxi reallocations. Specifically, rather than simply dividing the service area into grids,COXproposes a road connectivity aware clustering algorithm to divide the road network graph into zones for practical taxi dispatching. In addition,COXcomprehensively analyzes zone-level taxi demands and supplies through accurate taxi demand prediction and timely updates of taxi statuses.COXimproves the DRL modeling by integrating these derived contexts,e.g., state representation with complete demand/supply data and sequential action generation with full coordination among idle taxis. In particular, we implement an environment simulator to train and evaluateCOXusing a large real-world taxi dataset. Extensive experiments show thatCOXoutperforms state-of-the-art approaches on various performance metrics,e.g., on average improving the total order values by 6.74%, while reducing the number of unserved taxi orders and passengers’ waiting time by 4.92% and 44.84%, respectively.

Bus bunching, a phenomenon due to the failure of headway or timetable adherence, often causes low level of public transit service with poor bus on-time performance and excessive passenger waiting time. To mitigate bus bunching, an accurate and real-time prediction method plays an important role. In this paper, we propose a supply-demand seq2seq model called SD-seq2seq to predict bus bunching using smart card data. Features from both supply and demand sides of bus service are taken into account, like bus stop type, dwelling time, passenger demand and type, and so on. Extensive experiments on multiple bus routes in real world demonstrate that our method outperforms other baseline methods. The proposed method is expected to provide useful online information of bus operation to both bus operators and passengers.

Taxi ridesharing becomes promising and attractive because of the wide availability of taxis in a city and tremendous benefits of ridesharing, e.g., alleviating traffic congestion and reducing energy consumption. Existing taxi ridesharing schemes, however, are not efficient and practical, due to they simply match ride requests and taxis based on partial trip information and omit the offline passengers, who hail a taxi at roadside with no explicit requests to the system. In this paper, we consider the mobility-aware taxi ridesharing problem, and present mT- Share to address these limitations. mT-Share fully exploits the mobility information of ride requests and taxis to achieve efficient indexing of taxis/requests and better passenger-taxi matching, while still satisfying the constraints on passengers' deadlines and taxis' capacities. Specifically, mT-Share indexes taxis and ride requests with both geographical information and travel directions, and supports the shortest path based routing and probabilistic routing to serve both online and offline ride requests. Extensive experiments with a large real-world taxi dataset demonstrate the efficiency and effectiveness of mT-Share, which can response each ride request in milliseconds and with a moderate detour cost. Compared to state-of-the-art methods, mT-Share serves 42% and 62% more ride requests in peak and non-peak hours, respectively.

PageRank has a wide applications in online social networks and serves as an important benchmark to examine graph processing frameworks. Many efforts have been made to improve the computation efficiency of PageRank in shared-memory platforms, where a single machine can be sufficiently powerful to handle a large-scale graph. Existing methods, however, still suffer from synchronization issues and irregular memory accesses, which will deteriorate their overall performance. In this paper, we present an accelerated parallel PageRank computation approach, named APPR. By investigating the characteristics of parallel PageRank computation and degree distributions of social network graphs, APPR proposes a series of optimization techniques to improve the efficiency of PageRank computation. Specifically, a destination-centric graph partitioning scheme is designed to avoid the synchronization issues when concurrently updating the common vertex data. By exploiting power-law structure of social network graphs, APPR can intelligently schedule the computations of vertices to save computing operations. The vertex messages are adjusted by APPR for transmission to further improve the locality of memory accesses. Empirical evaluations are performed based on a set of large real-world graphs. Experimental results show that APPR significantly outperforms the state-of-the-art methods, with on average 2.4x speedup in execution time and 16.4x reduction in DRAM communication.

Crowdsensing tasks are usually described by certain features or attributes, and the task assignment essentially performs a matching with respect to the worker or user's preference on these features. However, the existing matching strategy could lead to a misaligned task coverage problem, i.e., some popular tasks tend to enter workers' candidate task lists, while some less popular tasks could be always unsuccessfully assigned. To ensure task coverage after the assignment, the system may have to increase their biding costs to reassign such tasks, which causes a high operational cost of the crowdsensing system. To address this problem, we propose to migrate certain qualified workers to the less popular tasks for increasing the task coverage and meanwhile, optimize other performance factors. By doing this, other performance factors, such as task acceptance and quality, can be comparably achieved as recent designs, while the system cost can be largely reduced. Following this idea, this article presents cTaskMat, which learns and exploits workers' task preferences to achieve coverage-ensured task assignments. We implement the cTaskMat design and evaluate its performance using both real-world experiments and data set-driven evaluations, also with the comparison with the state-of-the-art designs.

In this poster, we present our recent work, a wearable system for achieving real-time 3D arm skeleton. We have coped with the major challenge that the skeleton of each arm is determined from the locations of the elbow and wrist, whereas a wearable device only senses a single point from the wrist. Result shows that the potential solution space is huge. This underconstrained nature fundamentally challenges the achievement of accurate and real-time arm skeleton tracking. In this study, we propose Hidden Markov Model (HMM) state reorganization and hierarchical search two methods to improve the heavyweight computation of the state-of-art arm tracking model and achieve real-time tracking even on mobile phone.

This paper presents ArmTroi, a wearable system for understanding and analyzing the detailed arm motions of people primarily by using the motion sensors from wrist-worn wearable devices. ArmTroi can achieve real-time 3D arm skeleton tracking and reliable gesture inference tolerant to missing wearable sensors for enabling numerous useful application designs. We have coped with two major challenges through ArmTroi. First, the skeleton of each arm is determined from the locations of the elbow and wrist, whereas a wearable device only senses a single point from the wrist. We find that the potential solution space is huge. This underconstrained nature fundamentally challenges the achievement of accurate and real-time arm skeleton tracking. Second, wearable sensors may not reliably provide sensory data. For example, devices are not worn by the user, yet the learning tools for gesture inference, such as deep learning, typically have static network structures, which require nontrivial network adaptation to match the input's varying availability and ensure reliable gesture inference. We propose effective techniques to address above challenges, and all computations can be conducted on the user's smartphone. ArmTroi is thus a fully lightweight and portable system. We develop a prototype and extensive evaluation shows the efficacy of the ArmTroi design.

Accurate traffic prediction can benefit many smart city applications. Existing works mainly consider traffic prediction on each individual road segment, and heavily rely on some statistical or machine learning models, which suffer from either poor prediction accuracy or high computation overheads for predictions of the whole road network. In this paper, we instead consider the region-level traffic prediction that is still useful for many applications. To describe the regional traffic conditions and capture their spatio-temporal dependencies, we present a deep learning based model - DeepRTP. Specifically, we use a novel metric called Traffic State Index (TSI) to measure regional traffic conditions, and carefully classify traffic data into three categories that are used to capture hourly, daily, and weekly traffic patterns. Furthermore, we employ the convolutional and residual neural networks to model both spatial and temporal dependencies. Experimental results from real-world traffic data demonstrate that DeepRTP outperforms five baseline methods and can achieve higher prediction accuracy.

Mobile crowdsourcing (MCS) becomes an emerging paradigm for various useful urban sensing application designs by assigning the crowdsourcing tasks to the participants with rich-sensor equipped mobile devices. To effectively assign MCS tasks, many research efforts have been made in the literature. However, most prior schemes mainly optimize certain performance metrics in the assignment, yet overlooking other metrics, which thus cannot guarantee the overall system performance. This also limits the applicability of the proposed solution dedicated to the targeted performance metrics only. In this paper, we present UniTask, a unified task assignment design to address these issues. UniTask jointly considers the representative MCS performance metrics, including coverage, latency, and accuracy, to optimize the overall system utility. We mathematically formulate this problem and prove its NP-hardness. To efficiently schedule tasks, we also propose a utility-aware heuristic algorithm in UniTask. Moreover, a set of optimization techniques are further designed to enhance UniTask. Extensive evaluations are performed on two real-world datasets. Experimental results demonstrate that utility is an effective indicator of the system's overall performance. With an optimization on the system utility, UniTask can outperform the baseline methods on these individual performance metrics.

This paper presents a graph processing based traffic estimation system, GPTE, which is able to achieve high accuracy and high scalability to support city scale traffic estimation. GPTE benefits from its non-linear traffic correlation modeling and the graph-parallel processing framework built on clustered machines. By representing the road network as a property graph, GPTE decomposes the numerous computations involved in non-linear models to vertices and performs traffic estimation via neural network modeling and iterative information propagation. This paper presents our experiences in designing and implementing GPTE on top of the Spark, an emerging cluster computing framework. Extensive experiments are performed with real-world data input from Singapore's transport authority. Experimental results show that GPTE achieves as high as 88 percent accuracy in traffic estimation and up to 8× performance gain in computation efficiency with the optimization techniques applied. Comparison study demonstrates that GPTE outperforms the baseline solutions by 34 percent on accuracy and 46 percent on processing time.

Most digital maps are designed for vehicles and miss a great number of walkways that can facilitate people's daily mobility as pedestrians. Despite of such a fact, most existing map updating approaches only focus on the motorways. To fill the gap, this paper presents VitalAlley, a walkway discovery and verification approach with mobility data from large scale crowdsensing. VitalAlley aims to identify the uncharted walkways from the big but noisy personal mobility data and incorporate these findings into existing incomplete road maps. The implementation of VitalAlley faces the major challenges due to the unstructured nature of the walkways themselves and the noise from crowdsensing data. VitalAlley leverages different aspects of individual mobility to model and estimate the walkable areas, based on which representative walkways that connect known road segments or points of interest are extracted. To verify the new-found walkways, we further propose image based auto-verification with the help of publicly accessible street image database from Google Street View. VitalAlley is implemented and evaluated with real world crowdsensing data from the Singapore National Science Experiment. As a result, 736 walkways (totaling 161 km in distance) are identified from the mobility dataset collected from 108,337 students in Singapore. We manually verify 224 walkways totaling 32.4 km over a 9 km^2 district through on-site inspection. The results suggest over 96% accuracy of VitalAlley in discovering the walkways.

Existing digital maps mainly focus on motorways and miss many walkways facilitating people's daily mobility especially for pedestrians. Based on in-depth analysis of massive human mobility trajectories collected from the National Science Experiment (NSE) in Singapore, we propose a system, discovering walkways from the large scale crowdsensing mobility data. In this demo, we show the new-found walkways discovered by our system in Singapore through a custom visualization platform.

This paper presents a participatory sensing-based urban traffic monitoring system. Different from existing works that heavily rely on intrusive sensing or full cooperation from probe vehicles, our system exploits the power of participatory sensing and crowdsources the traffic sensing tasks to bus riders' mobile phones. The bus riders are information source providers and, meanwhile, major consumers of the final traffic output. The system takes public buses as dummy probes to detect road traffic conditions, and collects the minimum set of cellular data together with some lightweight sensing hints from the bus riders' mobile phones. Based on the crowdsourced data from participants, the system recovers the bus travel information and further derives the instant traffic conditions of roads covered by bus routes. The real-world experiments with a prototype implementation demonstrate the feasibility of our system, which achieves accurate and fine-grained traffic estimation with modest sensing and computation overhead at the crowd.

This paper presents a transport traffic estimation method which leverages road network correlation and sparse traffic sampling via the compressive sensing technique. Through the investigation on a traffic data set of more than 4400 taxis from Shanghai city, China, we observe nontrivial traffic correlations among the traffic conditions of different road segments and derive a mathematical model to capture such relations. After mathematical manipulation, the models can be used to construct representation bases to sparsely represent the traffic conditions of all road segments in a road network. With the trait of sparse representation, we propose a traffic estimation approach that applies the compressive sensing technique to achieve a city-scale traffic estimation with only a small number of probe vehicles, largely reducing the system operating cost. To validate the traffic correlation model and estimation method, we do extensive trace-driven experiments with real-world traffic data. The results show that the model effectively reveals the hidden structure of traffic correlations. The proposed estimation method derives accurate traffic conditions with the average accuracy as 0.80, calculated as the ratio between the number of correct traffic state category estimations and the number of all estimation times, based on only 50 probe vehicles' intervention, which significantly outperforms the state-of-the-art methods in both cost and traffic estimation accuracy.

This paper presents CSPR, a compressive-sensing-based approach for path reconstruction in wireless sensor networks. By viewing the whole network as a path representation space, an arbitrary routing path can be represented by a path vector in the space. As path length is usually much smaller than the network size, such path vectors are sparse, i.e., the majority of elements are zeros. By encoding sparse path representation into packets, the path vector (and thus the represented routing path) can be recovered from a small amount of packets using compressive sensing technique. CSPR formalizes the sparse path representation and enables accurate and efficient per-packet path reconstruction. CSPR is invulnerable to network dynamics and lossy links due to its distinct design. A set of optimization techniques is further proposed to improve the design. We evaluate CSPR in both testbed-based experiments and large-scale trace-driven simulations. Evaluation results show that CSPR achieves high path recovery accuracy (i.e., 100% and 96% in experiments and simulations, respectively) and outperforms the state-of-the-art approaches in various network settings.

This paper presents CSPR, a compressive sensing based approach for path reconstruction in wireless sensor networks. By viewing the whole network as a path representation space, an arbitrary routing path can be represented by a path vector in the space. As path length is usually much smaller than the network size, such path vectors are sparse, i.e., the majority of elements are zeros. By encoding sparse path representation into packets, the path vector (and thus the represented path) can be recovered from a small amount of packets using compressive sensing technique. CSPR formalizes the sparse path representation and enables accurate and efficient per-packet path reconstruction. CSPR is invulnerable to network dynamics and lossy links due to its distinct design. A set of optimization techniques are further proposed to improve the design. We evaluate CSPR in both testbed-based experiments and large-scale trace-driven simulations. Evaluation results show that CSPR achieves high path recovery accuracy (i.e., 100% and 96% in experiments and simulations, respectively), and outperforms the state-of-the-art approaches in various network settings.

Grouping sensor nodes with similar readings into the same cluster and scheduling them to alternately report their sensing readings is an effective and efficient method to perform approximate data collection, which exploits the tradeoff between data quality and energy consumption. However, to partition all sensor nodes into exclusive clusters while incurring as less communication overhead as possible in a distributed manner is a challenging task. In this paper, by exploiting inherent spatial and data correlation in wireless sensor network, we have proposed the distributed spatial correlation-based clustering algorithm to complete this tough mission. With nodes information exchange in certain region and the novel ranking strategy, our clustering algorithm can terminate in a small number of iterations. Extensive simulations show that the proposed algorithm outperforms three other noteworthy clustering algorithms, namely EEDC, ASAP and DClocal, on some key metrics, such as number of clusters, energy consumption for clustering, average dissimilarity of clusters and residual energy level of cluster heads.