In recent years, with the widespread application of robotics, designing efficient and generalizable robot control algorithms has become a critical research topic. Meta reinforcement learning as a method that enhances an agent's adaptability through cross-task learning, has demonstrated significant potential in robotic tasks. However, traditional meta reinforcement learning algorithms often rely on manually designed network architectures, which limit performance improvement and task adaptability. To address this issue, this paper proposes a meta reinforcement learning algorithm based on Neural Architecture Search. This method leverages robot simulation environments to automatically explore the optimal neural network architectures for meta reinforcement learning algorithms within the search space. By incorporating neural architecture search, the proposed algorithm aims to enhance their performance and efficiency. This study focuses on the design of the search space and search strategies tailored to meta reinforcement learning algorithms. Experimental results demonstrate that the proposed algorithm significantly improves task adaptability and training efficiency while reducing the labor costs associated with network architecture design, further demonstrating that this approach offers a new perspective for addressing robot control problems. It also establishes a solid foundation for future applications in real-world robotic systems.

Federated learning allows multiple users to collaboratively train models while protecting data privacy. However, for some users, the non-independent identically distributed nature of user data often reduces the accuracy of the global model. Existing personalized federated learning methods usually focus on individual users, which leads to problems such as bias and overfitting. This paper proposes a new distributed personalized federated learning framework based on Siamese neural networks (DPFL-SNN). First, a novel similarity calculation method is designed using the dual-branch structure of the Siamese neural network to effectively identify local users with similar data. Second, by combining this similarity calculation with blockchain technology, a new consensus algorithm is developed to achieve decentralization and reduce security risks. Simulations conducted on publicly available datasets demonstrate that the DPFL-SNN achieves higher accuracy compared to state-of-the-art personalized federated learning methods, thanks to enhanced collaboration among users with similar data.

Electroencephalogram (EEG)-based emotion recognition has achieved remarkable success in both subject-dependent and subject-independent scenarios. However, overcoming the challenges associated with reduced performance in EEG emotion recognition across devices, time, space, and subjects (i.e., cross-dataset) remains a significant obstacle for affective brain-computer interfaces (aBCIs). The key issue lies in the distributional mismatch between source and target domain EEG signals. To tackle the significant inter-domain differences in cross-dataset EEG emotion recognition, this paper introduces an innovative framework termed the Emotional EEG Style Transfer Network (E$^{2}$STN), which aims to effectively capture the emotional content information from the source domain and the style features from the target domain, facilitating the reconstruction of stylized emotion EEG representations. These stylized EEG representations significantly enhance the discriminative prediction performance in cross-dataset EEG emotion recognition. Specifically, E$^{2}$STN consists of three key modules: a Transfer Module for domain style transfer, a Transfer Evaluation Module for evaluating transfer quality, and a Discriminative Module for making discriminative predictions. Extensive experiments demonstrate that E$^{2}$STN achieves the state-of-the-art performance in cross-dataset emotion EEG recognition. To the best of our knowledge, this is the first work to explicitly address cross-dataset emotion EEG recognition. The experimental results provide a valuable benchmark for future research in this area.

Fine-grained urban traffic data are often incomplete owing to limitations in sensor technology and economic cost. However, data-driven traffic analysis methods in intelligent transportation systems (ITSs) heavily rely on the quality of input data. Thus, accurately estimating missing traffic observations is an essential data engineering task in ITSs. The complexity of underlying node-wise correlation structures and various missing scenarios presents a significant challenge in achieving high-precision estimation. This study proposes a novel multiview neural network termed MVCV-Traffic, equipped with a cross-view learning mechanism, to improve traffic estimation. The contributions of this model can be summarized into two parts: multiview learning and cross-view fusing. For multiview learning, several specialized neural networks are adopted to fit diverse correlation structures from different views. For cross-view fusing, a new information fusion strategy merges multiview messages at both feature and output levels to enhance the learning of joint correlations. Experiments on two real-world datasets demonstrate that the proposed model significantly outperforms existing traffic speed estimation methods for different types and rates of missing data.

The accurate discovery of disease-outbreak source areas plays a critical role in the effective containment of epidemics at an early stage. Existing relevant methods are mostly implemented by tracing the source of disease outbreaks directly from the distribution of confirmed cases in the Euclidean space. In reality, in most respiratory infectious diseases, crowd gathering caused by resident trips significantly increases the risk of exposure to potentially infected persons, making it the driving force behind the outbreak of the disease. In light of this, this study proposes a network-constrained ring-shaped hotspot detection method based on the classical spatial scan statistic model. This new method can be used to determine the sizes of the scanning window in an adaptive manner by using the information of the trip distance distributions. Considering the centered traffic analysis zone in each hotspot as a candidate outbreak source area, a multi-factor coupling model was designed by characterizing both the areal vibrancy and resident trip distributions to further identify the potential disease-outbreak source areas. A case study on the outbreaks of COVID-19 in Wenzhou, China, was carried out to evaluate the practicability and effectiveness of the proposed method. The comparison results demonstrate the effectiveness of the proposed method for detecting areas and sources of outbreaks.

The implementation of a variety of complex and energy-intensive mobile applications by resource-limited mobile devices (MDs) is a huge challenge. Fortunately, mobile edge computing (MEC) as a new computing paragon can offer rich resources to perform all or part of the MD’s task, which greatly reduces the energy consumption of the MD and improves the quality of service (QoS) for applications. However, offloading tasks to the edge server is vulnerable to attacks such as tampering and snooping, resulting in a deep learning (DL) security feature developed by major cloud service providers. An effective security strategy method to minimize ongoing attacks in the MEC setting is proposed. The algorithm is based on the synthetic principle of a special set of strategies, and it can quickly construct suboptimal solutions even if the number of targets achieves hundreds of millions. In addition, for a given structure and a given number of patrollers, the upper bound of the protection level can be obtained, and the lower bound required for a given protection level can also be inferred. These bounds apply to universal strategies. By comparing with the previous three basic experiments, it can be proved that our algorithm is better than the previous ones in terms of security and running time.

The Internet of Things (IoT) interconnects various devices and services, of which recommendation services are an important component to help the development of IoT applications. Furthermore, without the aid of suitable online recommendation systems, Internet users will be overwhelmed by the tremendous amount of contents. Researchers have thus developed a large volume of recommendations. However, they are all flawed with high complexity, cold start issues, inability to generalize, etc. In recent years, some researchers had turned to variational inference (VI)-based recommendation systems, which can solve the above problems to some extent. However, these VI-based recommendations are merely hybrid methods of VI with the existing recommendation algorithms and are unable to be implemented well in real practices. Therefore, developing algorithms that can overcome these limitations of the existing online recommendation systems is essential for convenient and useful Internet searches. In this paper, we propose, develop, implement and test a more general, new and innovative Stein Variational Recommendation System algorithm (SVRS) to tackle the long plaguing recommendation problems. Based on Stein’s identity, the SVRS algorithm can compute the feature vector of existing users and items it had rated, and further predict the ratings for users that have not been engaged with certain content. SVRS provides more general insights into the forming of user ratings, can be easily extended to higher dimensions and has the merits of low complexity, easy scaling and generalizability. Experiments show that SVRS outperforms the other existing type of recommendation algorithms and it has higher accuracy in terms of mean absolute error (MAE) and root mean square error (RMSE).

This paper proposes an offloading strategy for communication resources and computing resources. Based on the communication background of NOMA, this strategy combines D2D communication and cellular mobile communication to consider the use of resources in the communication process. According to the performance of communication resources, the power of cellular network users and D2D users is controlled to make the task unloading decision between users and small MEC server base stations. The purpose is to find the minimum sum of the total cost of the user group within the coverage of the entire base station. The cost is obtained from the task completion time and the base station usage payment. The experimental results show that this method can find the optimal solution of the user group cost, and it also has a good performance in user fairness. In reality, taking D2D communication into consideration for coordinated distribution in a cellular network can greatly reduce the cost of the two.

In order to solve the problem of task offloading with dependent subtasks in mobile edge computing (MEC), we propose a graph mapping offloading model based on deep reinforcement learning (DRL). We model the user's computing task as directed acyclic graph (DAG), called DAG task. Then the DAG task is converted into a topological sequence composed of task vectors according to the custom priority. And the model we proposed can map the topological sequence to offloading decisions. The offloading problem is formulated as a Markov decision process (MDP) to minimize the trade-off between latency and energy consumption. The evaluation results demonstrate that our DRL-based graph mapping offloading model has better decision-making ability, which proves the availability and effectiveness of the model.

Sparse Mobile Crowdsensing (SparseMCS) has become an efficient method for collecting and inferring sensor data at the urban scale, attracting many researchers' attention. However, there is a risk of location privacy being leaked when the participants complete the task. The existing SparseMCS location privacy-preserving mechanisms focus on protecting location information from the spatial dimension while ignoring the time dimension's location privacy. To solve this problem, we combine spatiotemporal activity privacy with location differential privacy and propose a novel location privacy-preserving mechanism. More specifically, location differential privacy is used to limit the adversary's information gain in the space dimension, and spatiotemporal activity privacy is used to determine the adversary's information gain in the time dimension. The combination of the two can protect the user's spatiotemporal activity patterns over a period. To achieve efficient sensing data collection, we train the privacy-preserving model when the server is offline. Since the location privacy-preserving mechanism will affect the accuracy of the sensing data, in order to reduce the data loss, we design a privacy-preserving model consisting of three parts: First, to fit the original location sensing data to the confusion location, we learned a data adjustment function. Secondly, we solve the optimal position confusion function through nonlinear programming that aims to minimize data loss. Third, we use an inference algorithm that is sensitive to data loss to improve the accuracy of the final inferred data. Our work is within the acceptable error range of 10-3~10-2, which can achieve more comprehensive and stronger location privacy.

Abstract In recent years, the success of deep learning in natural scene image processing boosted its application in the analysis of remote sensing images. In this paper, we applied Convolutional Neural Networks (CNN) on the semantic segmentation of remote sensing images. We improve the Encoder- Decoder CNN structure SegNet with index pooling and U-net to make them suitable for multi-targets semantic segmentation of remote sensing images. The results show that these two models have their own advantages and disadvantages on the segmentation of different objects. In addition, we propose an integrated algorithm that integrates these two models. Experimental results show that the presented integrated algorithm can exploite the advantages of both the models for multi-target segmentation and achieve a better segmentation compared to these two models.

This article studies the attacker identification issue in the Industrial Internet of Things (IIoT). There have been already some work that uses device fingerprinting to identify attackers, and the transmission offset of the device internal clock signals is used as the device's fingerprint. However, the existing work to measure the offset relies on the periodic transmission of signals, but in many types of IIoT devices, the signal transmission is aperiodic. To eliminate the limitation on the periodicity, in this article, we design an algorithm, Fiden, to fingerprint heterogeneous IIoT devices without considering the periodicity. This algorithm extracts the patterns from the time series of signal transmission, and then learns the fingerprint of a device by clustering the patterns. We demonstrate the applicability of Fiden by a real case study on the communications environment of the vehicle industry. The results show that the proposed algorithm helps identify the devices-mounted attacks. Compared with the clock-based intrusion detection system (CIDS), when the timestamp and accumulated clock offset of the signal transmission are used as the features for pattern extraction, Fiden's accuracy is increased by 15% and 15% to these two features, precision is increased by 25% and 23%, recall is increased by 28% and 18%, and F1 score is increased by 28% and 21%.

Vehicular ad hoc networks (VANETs) has become an important part of modern intelligent transportation systems (ITS). However, under the influence of malicious mobile vehicles, offloading vehicle tasks to the cloud server is threatened by security attacks. Edge cloud offloading (ECCO) has considered a promising approach to enable latency-sensitive VANET. How to solve the complex computation offloading of vehicles while ensuring the high security of the cloud server is an issue that needs urgent research. In this paper, we studied the safety and offloading of multi-vehicle ECCO system based on cloud blockchain. First, to achieve consensus in the vehicular environment, we propose a distributed hierarchical software-defined VANET (SDVs) framework to establish a security architecture. Secondly, to improve the security of offloading, we propose to use blockchain-based access control, which protects the cloud from illegal offloading actions. Finally, to solve the intensive computing problem of authorized vehicles, we determine task offloading via jointly optimizing offloading decisions, consensus mechanism decisions, allocation of computation resources and channel bandwidth. The optimization method is designed to minimize long-term system of delays, energy consumption, and flow costs for all vehicles. To better resolve the proposed offloading method, we develop a new deep reinforcement learning (DRL) algorithm via utilizing extended deep Q-networks. We evaluate the performance of our framework on access control and offloading through numerical simulations, which have significant advantages over existing solutions.

In the era of big data, the same question can get many answers from multiple sources. These answers may conflict with each other. Therefore, how to get the true information (i.e., the truths) from many answers has been a hot research topic. At present, there are many truth discovery methods which employ source reliability to improve the quality of truths. Most existing methods can only handle the categorical data or numerical data, while performs bad on text data. Meanwhile, we observed that the text data contains not only the question answer, but also some implicit information, for example, possible, may, make sure, similar, the same as, etc. These words have nothing to do with the answer but can reflect the self-confidence degree of source. In this paper, we propose a truth discovery framework which takes the implicit information into account. We first analyze text data and extract the answers of questions, then we create two dictionaries composed of self-confidence increasing words and self-confidence decreasing words respectively. Using the dictionaries we extract the self-confidence information from answer descriptions. Finally, we take full advantage of the self-confidence information to improve the performance of truth discovery. We perform experiments using a categorical data and a real-world Chinese text data. Comparing with other methods, our framework performs better, which demonstrates the superiority of our proposed framework.

With the fast growing of artificial intelligence (AI), more and more applications require querying uncertain data, especially from social media and crowd sourcing platform. In situations where it is impossible to increase data quality by controlling the sources, we may resort to algorithms to make the best use of the collected data. Since crowdsourcing provides a useful way to distributing tasks to mass people, and collects labels from as many workers as possible, many researchers have been study crowd-sourcing inference algorithms. In our work, we propose a novel crowd-sourcing inference algorithm to infer ground truth and obtain worker reliability and task difficulty at the same time.

We propose an algorithm for constructing efficient security strategies in the mobile edge computing (MEC), where the protected targets are nodes connected to the MEC and the mobile users (MUs) are agents capable of preventing undesirable activities on the nodes. The algorithm is designed based on the synthetic principles of a specific set of strategies, and it can quickly construct suboptimal solutions even if the number of targets reaches hundreds of millions.

Spatiotemporal co-occurrence patterns (STCOPs) are subsets of Boolean features whose instances frequently co-occur in both space and time. The detection of STCOPs is crucial to the investigation of the spatiotemporal interactions among different features. However, prevalent STCOPs reported by available methods do not necessarily indicate the statistically significant dependence among different features, which is likely to result in highly erroneous assessments in practice. To improve the reliability of results, this paper develops a statistical method to detect STCOPs and discern their statistical significance. The proposed method detects STCOPs against the null hypothesis that the spatiotemporal distributions of different features are independent of each other. To construct the null hypothesis, suitable spatiotemporal point-process models considering spatiotemporal autocorrelation are employed to model the distributions of different features. The performance of the proposed statistical method is assessed by synthetic experiments and a case study aimed at identifying crime patterns among multiple crime types in Portland City. The experimental results demonstrate that the proposed method is more effective for detecting meaningful STCOPs than the available alternative methods.

Simultaneous multithreading (SMT) is a unique computer architecture feature to increase the pipeline utilization and therefore, increase the instruction throughput. It improves instruction level throughput by simultaneously filling both vertical and horizontal super-scalar pipeline slots that are left unfilled by native threads. So far SMT has been implemented in hardware. However, hardware SMT implementations have its limitations. First, it is complex and expensive to implement—only higher-end processors are equipped with it even though lower-end processors have same pipeline design as the higher-end variants and might also benefit from it. SMT also introduces great power/energy and area overheads. Moreover, being a hardware feature, SMT is limited by the range and depth of instruction analysis that it can afford at execution time, therefore it is unlikely to benefit from high-level software knowledge about instruction mix and might lose many improvement opportunities. In this paper, we address the limitation of the hardware-based SMT and introduce CSSMT: Compiler based Software Simultaneous Multithreading (SMT). The main contribution of CSSMT is that it exploits high- level program profiles to purposefully "re-mix" instructions from multiple programs to better fill vertical and horizontal super- scalar pipeline slots so that the overall throughput is improved. Furthermore, CSSMT is a software-transformation technique that enables SMT at software level during compilation time. Therefore, it can help overcome the limitation of the hardware-based SMT implementation and is more portable. We test CSSMT with programs from SPEC2006 and NAS benchmarks and achieve up to 12% speedup of execution time (30.7% improvement in terms of multi-program throughput).

This paper studies the secure transmission rate issue between sensors and the remote controller to defend the jamming and eavesdropping attacks in green cyber-physical transportation systems. In this system, the traffic sensor transmits the transportation state information to the remote controller via wireless networks. Due to the broadcast characteristics of the wireless communication, the systems are vulnerable to the eavesdropping and jamming attacks. In this paper, we study how to maximize the secure transmission rate between sensors and the controller in the presence of a malicious eavesdropper and a jammer. Specifically, the malicious jammer is smart and can choose the optimal power strategy to maximize the side effect with the knowledge of sensor's transmission power. For the purpose of achieving the maximum utility, the optimal strategy is determined via adjusting the sensor's transmission power according to the control feedback conditions. We consider the single-antenna model and the multiantenna model to formulate this problem as an optimization problem based on a Stackelberg game. We then prove the existence of the Stackelberg equilibrium via the interaction between the sensor and the jammer. Moreover, we present two algorithms to obtain the optimal transmission strategy, i.e., a stochastic algorithm with feedback and renewed intelligent simulated annealing. Finally, extensive simulations and trace experimental results are presented to verify our theoretical analysis.

Traffic flow prediction is an important research issue to avoid traffic congestion in transportation systems. Traffic congestion avoiding can be achieved by knowing traffic flow and then conducting transportation planning. Achieving traffic flow prediction is challenging as the prediction is affected by many complex factors such as inter-region traffic, vehicles' relations, and sudden events. However, as the mobile data of vehicles has been widely collected by sensor-embedded devices in transportation systems, it is possible to predict the traffic flow by analysing mobile data. This study proposes a deep learning based prediction algorithm, DeepTFP, to collectively predict the traffic flow on each and every traffic road of a city. This algorithm uses three deep residual neural networks to model temporal closeness, period, and trend properties of traffic flow. Each residual neural network consists of a branch of residual convolutional units. DeepTFP aggregates the outputs of the three residual neural networks to optimize the parameters of a time series prediction model. Contrast experiments on mobile time series data from the transportation system of England demonstrate that the proposed DeepTFP outperforms the Long Short-Term Memory (LSTM) architecture based method in prediction accuracy.

This study proposes an eco-friendly navigation algorithm, eRouting, to save energy and reduce CO2emission. The important research issue of traffic information industry, eco-friendly navigation, has been widely studied. As an improvement, in this paper, combining real-time traffic information and a representative factor-based energy/emission model, a calculated route is dynamically adjusted during the travel of a vehicle. eRouting is a centralized algorithm. It profits from the following aspects to achieve improved performance: a representative factor-based energy/emission model, a real-time traffic information-based dynamic adjustment, and an objective function to control the optimization direction to optimize the final energy consumption of vehicle's travel. As a peculiarity of this paper, the design of the representative factor-based model mines the impact of road-level parameters on energy consumption and CO2emission. Such mining is helpful to improve the pertinence of a model by formulating the key influence factors into the model. Experimental results are presented to prove the validity of eRouting. In addition, by contrast experiments, eRouting shows improved performance compared with an eco-friendly navigation algorithm and three traditional navigation algorithms.

In recent years, WiFi fingerprint-based localization has received much attention due to its deployment practicability. Although existing works show WiFi fingerprinting can achieve good localization accuracy, the experiments were conducted under their own testbeds within a small area and a short period. In this work, we investigate the impact of different indoor environmental factors, such as temporal and spatial similarity, on the performance of WiFi fingerprinting. We find that, WiFi fingerprinting is highly environment-sensitive. In an open space, it is quite challenging to find spatially varying but temporally stable signatures for adjacent reference locations. To address this issue, we propose a heatmap-based WiFi fingerprinting (called HMF) by utilizing layout construction as an additional input to improve WiFi fingerprint localization in open space environment. Our experimental results show, HMF can improve existing WiFi fingerprinting schemes like Radar and Horus by 28% and 80% in moderately open space, e.g., a wide corridor.

In this article, we address a sleep scheduling scheme that ensures a coverage degree requirement based on the dangerous levels of the toxic gas leakage area, while maintaining the global network connectivity with minimal awake-nodes compared to other schemes that increase the number of awake- nodes over the entire network.

In this article, we propose a new scheme DeGas to determine the boundary area of the toxic gas with planarization algorithm. This detected boundary area will ensure safe area around the toxic gas and will provide a decision reference for evacuation and rescue of the first-line workers in the large-scale petrochemical plants. Understanding the implications of these observations enable to find out an optimal trade-off between the cost of a number of deployed sensor nodes and the accuracy of the estimated toxic gas boundary area size.

Understanding the propagation dynamics of information/an epidemic on complex networks is very important for discovering and controlling a terrorist attack, and even for predicting a disease outbreak. As an effective method, with analyzing the structure of a propagation network, a large number of previous studies have analyzed the propagation dynamics. Most of these studies are based on a special network structure to make such analysis. However, a propagation network has dynamically changed structure during the propagation. How to track, recognize and model such dynamic change is a big challenge. Along with the popularity of smart devices and the rapid development of the Internet of Things (IoT), massive mobile data is automatically collected. In this article, as a typical use case, we investigate the impact of network structure on epidemic propagation dynamics by analyzing the massive mobile data collected from smart devices carried by the volunteers of Ebola outbreak areas. From this investigation, we obtain two observations. Based on these observations and the analytical ability of Apache Spark on streaming data and graphs, we propose a simple model to track and recognize the dynamic structure of a network. Moreover, we introduce and discuss open issues and future work for developing this proposed recognition model.

Industrial Wireless Sensor Networks (IWSNs) are expected to offer promising monitoring solutions to meet the demands of monitoring applications for fault diagnosis in large-scale petrochemical plants, however, involves heterogeneity and Big Data problems due to large amounts of sensor data with high volume and velocity. Cloud Computing is an outstanding approach which provides a flexible platform to support the addressing of such heterogeneous and data-intensive problems with massive computing, storage, and data-based services. In this paper, we propose a Cloud-based Data-intensive Framework (CDF) for on-line equipment fault diagnosis system that facilitates the integration and processing of mass sensor data generated from Industrial Sensing Ecosystem (ISE). ISE enables data collection of interest with topic-specific industrial monitoring systems. Moreover, this approach contributes the establishment of on-line fault diagnosis monitoring system with sensor streaming computing and storage paradigms based on Hadoop as a key to the complex problems. Finally, we present a practical illustration referred to this framework serving equipment fault diagnosis systems with the ISE.

Threads usually do not use all available resource son a processor at the same time. For example, some program regions are dominated by memory access instructions while leaving computing units idling, or on the contrary are busy calculating but not using memory I/O. This observation points to an optimization opportunity that if the overall system throughput is the goal, one thread can "piggyback" on another thread without significantly hurting the other thread's performance if the first thread only uses CPU resources that are not heavily used by the other. Clearly the "piggybacking" opportunities are an instruction-level mechanism and OS-level resource management cannot efficiently leverage them. In this paper we propose a new compiler transformation that merges two or multiple program sat instruction level to take advantage of the resource piggybacking opportunities. In particular, we propose program profiling techniques to identify good merging candidates and develop transformation to merge the programs while maintain their original semantics. We evaluated our approach on synthesized benchmarks and 3 NPB benchmarks. In one case, we achieve50% speedup over running the pre-merged benchmarks serially, or 30% speedup over running them in parallel and letting OSmanage their execution.

In existing WiFi-based localization methods, smart mobile devices consume quite a lot of power as WiFi interfaces need to be used for frequent AP scanning during the localization process. In this work, we design an energy-efficient indoor localization system called ZigBee assisted indoor localization (ZIL) based on WiFi fingerprints via ZigBee interference signatures. ZIL uses ZigBee interfaces to collect mixed WiFi signals, which include non-periodic WiFi data and periodic beacon signals. However, WiFi APs cannot be identified from these WiFi signals by ZigBee interfaces directly. To address this issue, we propose a method for detecting WiFi APs to form WiFi fingerprints from the signals collected by ZigBee interfaces. We propose a novel fingerprint matching algorithm to align a pair of fingerprints effectively. To improve the localization accuracy, we design the K-nearest neighbor (KNN) method with three different weighted distances and find that the KNN algorithm with the Manhattan distance performs best. Experiments show that ZIL can achieve the localization accuracy of 87%, which is competitive compared to state-of-the-art WiFi fingerprint-based approaches, and save energy by 68% on average compared to the approach based on WiFi interface.

Location-Based services with social networks improve users' experience and enrich people's social live. However, location information is often inadequate due to privacy and security concerns. We seek to infer users' ‘Current City’ on Facebook for coarse location based applications. We first extract users' multiple explicit and implicit location attributes, and analyze correlations of these attributes from two perspective: user-centric and user-friends. We observe that both user-centric and user-friends location attributes tightly correlate to a user's Current City (e.g., 60% of users stay in their hometown, 60% of users live in the same city as 50% of their friends). Based on extensive analysis and observations on location attributes correlations, we have constructed a Current City Prediction model (CCP) using artificial neural network (ANN) learning frameworks. The experimental results indicate that we achieve accuracy levels of 84% for city-level prediction and 98% for country-level which are increases of 9% and 18%, respectively than what is possible with Tweecalization.

Finding basic laws that govern human crowd behavior is a subject that deserves to study. Crowd behavior is a natural instinct for human, which directly impacts how we form opinions and make decisions. Moreover, it is common that people change their behavior in a group. In pervasive computing research, substantial work has been directed towards discovering human movement patterns based on wireless networks. This research has, however, mainly been focused on movements of individuals. Mobile phones offer on-body tracking and they are already deployed on a large scale, allowing the characterization of user behavior through the information related to individual movements. In this paper, we observe and analyze the impact of friendship and location attributes on crowd behavior, using location-based wireless mobility information. These preliminary studies will be a good cornerstone for a crowd behavior prediction.

In this poster, we propose a novel energy efficient asynchronous MAC protocol for WSNs-WX-MAC. WX-MAC supports nodes to exchange their sampling schedules by query and report mechanisms. With the aid of neighbors' information and strobed preamble approach, WX-MAC shortens senders' preamble before data transmission and reduces nodes' wakeup duration. WX-MAC not only reduces the energy consumption on the senders and targets, but also decreases the possibility of overhearing on the surrounding neighbors, leading to a further energy conservation. Preliminary simulations have shown that WX-MAC saves more energy than B-MAC, Wise MAC, and X-MAC.

Most indoor localization algorithms are based on Received Signal Strength (RSS), in which RSS signatures of an interested area are annotated with their real recorded locations. However, according to our experiments, RSS signatures are not suitable as the unique annotations (like Fingerprints) of recorded locations. In this study, we investigate the characteristics of RSS (e.g., how the RSS values change as time goes on and between consecutive positions?). On this basis, we design LuPI (Locating using Prior Information) that exploits the characteristics of RSS: with user motion, LuPI uses novel sensors integrated in smartphones to construct the RSS variation space (like radio map) of a floor plan as prior information. The deployment of LuPI is easy and rapid since little human intervention is needed. In LuPI, the calibration of ``radio map'' is crowd-sourced, automatic and scheduled. Experimental results show that LuPI achieves comparable location accuracy to previous approaches, even without the statistical information of site survey.

SUMMARY Wireless converge‐cast networks (WCNs), such as data collection‐based wireless sensor networks, exhibit certain phenomena called funneling effect, where the region close to the sink node is heavily overloaded. In this paper, we identify that the funneling effect occurs not only close to the sink but also within the network region where nodes have collision and induce heavy traffic to relay; we name it hot‐spot funneling effect. This paper aims to improve the throughput and fairness of WCNs by mitigating the micro funneling effect. We propose a new mechanism, the backoff differentiation for contention resolution (BDCR), which is targeted to a system‐wide high throughput on the basis of the contention resolution mechanism. To achieve high spatial reuses, BDCR divides the network into several regions and does backoff differentiation within each region. Within each backoff differentiation region, the backoff window range is adjusted according to the traffic rate, and at the same time, the backoff values are set with the awareness of the traffic intensity level. All regions share the same algorithm, which uses Kelly's rate control theory and method to allow each sensor to locally adjust its backoff value. One of the key advantages of BDCR is that it is extremely easy to implement. With extensive simulations and testbed experiments, BDCR is proved to achieve much higher throughput over the traditional carrier sense multiple access and some recent media access control protocols in literature, particularly when the network suffers intensive congestions. Copyright © 2012 John Wiley & Sons, Ltd.

Data aggregation, as a primitive communication task in wireless networks, can reduce the communication complexity. However, in-network aggregation usually brings an unavoidable security defect. Some malicious nodes may control a large percentage of the whole network data and compel the network misbehave in an arbitrary manner. Thus, locating the malicious nodes to prevent them from further disaster is a practical challenge for data aggregation schemes. Based on the grouping and localization techniques, we propose a novel integrated protocol to locate malicious nodes. The proposed protocol does not rely on any special hardware and requests only incomplete information of the network from the security schemes. We also conduct simulation study to evaluate the proposed protocol.

Parallelism helps performance but at the same time stresses computer resources that are shared among threads. In this paper, we propose a low-overhead hardware counter based profiling method to accurately identify time-relevant contention locations in the program, then these contentions are mitigated so that performance of multithreading tasks can be boosted by the reduction of unnecessary contention cycles. In our preliminary experiment using NAS Parallel Benchmark (NPB), the contention searching algorithm is able to find an severe memory contention loop in FT code. After contention mitigation, more than 10% of the total cycles is eliminated, and the execution time of FT is reduced by 3% at the same time.

Location information of nodes is the basis for many applications in wireless sensor networks (WSNs). However, most previous localization methods make the unrealistic assumptions: (i) all nodes in WSN are always awake and (ii) the radio range of nodes is an ideal circle. This overlooks the common scenario that sensor nodes are duty-cycled in order to save energy and the radio range of nodes is irregular. In this paper we revisit the Distance-Vector-based (DV-based) positioning algorithms, particularly, Hop-Count-Ratio based Localization (HCRL) algorithm and investigate the following problems: (i) how is the relationship between the number of sleeping neighbor sensor nodes and the localization accuracy and (ii) how is the relationship between the degree of irregularity (DOI, which is a parameter of radio range irregularity) and the localization accuracy. We conduct a large number of experiments in WSNs' simulator NetTopo, and find that the parameters: the number of waking nodes, DOI, anchor node density and localization error, are interactional, i.e., for a given deployed static WSN, there is an optimal number of waking nodes and an optimal anchor node density, which can minimize network energy consumption without losing much of the localization accuracy. Furthermore, waking up more sensor nodes cannot always help to increase the localization accuracy, which actually is different from our intuitive thinking: more waking nodes can help to increase the localization accuracy of DV-based localization algorithms at all time.

In Wireless Sensor Networks (WSNs) a multitude of location-dependent applications have been proposed recently, which is very intriguing for researchers to discover and design more accurate and cost-effective localization algorithms. In an isotropic networks, the Euclidean distance between a pair of nodes may not correlate closely with the hop count between them because the corresponding shortest path may have to curve around intermediate holes, resulting in poor distance estimation. And without the help of a large number of uniformly deployed seed nodes, those schemes fail in an isotropic WSNs. To address this issue and improve the accuracy of localization, we propose the Removing Heavily Curved Path (RHCP) scheme in this paper. RHCP takes advantage of selecting the paths which are not heavily affected by the holes to recalculate the location of each unknown node. Through simulation, the results reveal that RHCP performs better than original DV-Hop in an isotropic networks with different shape of holes. In addition, through iterations of RHCP, the results get improved for different anchor node densities.

Range-Free algorithms, appealing to people for their cost-efficiency, suffer from the precision problem. Some methods try to combine received signal strength indication (RSSI) with range-free localization algorithms to improve the accuracy, but RSSI is sensitive to the radio irregularity. Based on the well known RIM model, we present a new method of RSSI calibration, namely MRIRC, to mitigate the impact of radio irregularity. MRIRC divides nodes within a continuous angle into groups with the same level of RSSI deviation. By doing this, given an irregular deviation input, MRIRC can get a maximum angle (worst case), which guarantees that the nodes in the same group are in the same level of radio irregularity, thereby improving the accuracy of the distance estimations. We conduct simulations for large-scale sensor networks, and the results show that MRIRC achieves superior performance over the other two typical Range-Free algorithms.

Wireless sensor networks (WSNs) have been widely used in many applications. The highly correct location information of sensor nodes is crucial for these applications. Nowadays there are mainly two types of localization algorithms: Range-based localization algorithms and Range-free localization algorithms. The drawback of former type is strict requirements on the hardware configuration. The latter type is cost-effective alternative approach. However, this approach can achieve high accuracy only in some ideal scenarios, and some methods even require a lot of complex calculations. In this paper, we put forward a novel PAV-based Calibration localization method (PAVC) locating the unknown nodes. The Pool Adjacent Violators (PAV) algorithm can be used to calibrate the average distance per hop of hop-based positioning algorithm. PAVC improves the localization accuracy compared with previous hop-based algorithms, which is demonstrated by the simulation and testbed results.

This paper aims at mitigating the so-called Funneling Effect for S-MAC, particularly by improving the throughput and fairness of S-MAC. Wireless sensor networks (WSNs) exhibit some phenomenon named Funneling Effect resulting from the accumulation of disproportionate large number of packets in the regions close to the sink. The collision and congestion due to the Funneling Effect strongly weaken the vitality and robustness of WSNs. As for S-MAC which achieves great energy efficiency, the mitigation of funneling effect seems more significant and urgent. In this paper, targeted to alleviate the funneling effect for S-MAC, we propose a SMAC-based proportional fairness backoff scheme (SPFB). Based on the schedule and contention scheme in S-MAC, SPFB employs Kelly's shadow price theory to achieve the proportional fairness as well as optimizes the back off mechanism to improve the throughput. The contention window range is dynamically adjusted according to the load of individual node. With extensive simulations, we can show that SPFB can achieve much higher throughput than traditional S-MAC, especially when the network is heavy loaded. SPFB can also gain good energy efficiency.

Localization problem in wireless sensor networks (WSNs) has been widely studied recently. However, most previous work simply assume that all the nodes stay awake during the localization phase. This assumption clearly overlooks the common scenario that sensor nodes are usually duty-cycled in order to save energy. In this paper we propose a kind of novel DV (distance vector)-based localization algorithm which performs pretty good in duty-cycled network. In order to get a good localization accuracy, the DV-based positioning algorithms need to keep a critical minimum average neighborhood size (CMANS) for every sensor node. However, in the time-varying connectivity (TVC) (this phenomenon results from duty-cycling) network, it is difficult to keep CMANS for every node all the time. We can use CKN sleep scheduling algorithm to tackle this problem. CKN sleep scheduling algorithm can save energy while keeping certain CMANS. We further propose a novel localization algorithm: Spline Function Localization (SFL) algorithm which guarantees high accuracy even under small neighborhood size. Finally, we estimate the performance of our algorithm and compare with several classical DV-based localization algorithms (DVHOP and HCRL (Hop-Count-Ratio based Localization)) in simulation. Experimental results confirm that our algorithm has much higher accuracy under duty-cycled network.

This paper aims at improving the throughput of the wireless sensor networks (WSNs), particularly to overcome the so-called funneling effect for WSNs with converge-cast patterns. Due to the disproportionate larger number of packets accumulated in the sensors that are closer to the sink, there is a need to decrease the collisions and increase the throughput around the sink area as well as the nodes that experience a heavy pass-through traffic. In this paper, we proposed a new scheme, namely PFB (Proportional Fairness Backoff), which provides additional scheduling opportunities to nodes closer to the sink. The new scheme employs Kelly's shadow price theory to achieve the proportional fairness, which takes advantage of the tree topology that is the de facto standard in today's WSNs. In PFB, the size of backoff window is dynamically adjusted with respect to the height of nodes belong in the tree. With close-form analysis and extensive simulations, we show that PFB can achieve up to 100% throughput increase over the widely used CSMA when the network is highly loaded.