Search

In this letter, we consider a distributed submodular maximization problem for multi-robot systems when attacked by adversaries. One of the major challenges for multi-robot systems is to increase resilience against failures or attacks. This is particularly important for distributed systems under attack as there is no central point of command that can detect, mitigate, and recover from attacks. Instead, a distributed multi-robot system must coordinate effectively to overcome adversarial attacks. In this work, our distributed submodular action selection problem models a broad set of scenarios where each robot in a multi-robot system has multiple action selections that may fulfill a global objective, such as exploration or target tracking. To increase resilience in this context, we propose a fully distributed algorithm to guide each robot's action selection when the system is attacked. The proposed algorithm guarantees performance in a worst-case scenario where up to a portion of the robots malfunction due to attacks. Importantly, the proposed algorithm is also consistent, as it is shown to converge to the same solution as a centralized method. Finally, a distributed resilient multi-robot exploration problem is presented to confirm the performance of the proposed algorithm.

We introduce and study the problem of planning a trajectory for an agent to carry out a scouting mission while avoiding being detected by an adversarial opponent. This introduces a multi-objective version of classical visibility-based target search and pursuit-evasion problem. In our formulation, the agent receives a positive reward for increasing its visibility (by exploring new regions) and a negative penalty every time it is detected by the opponent. The objective is to find a finite-horizon path for the agent that balances the trade off between maximizing visibility and minimizing detectability. We model this problem as a discrete, sequential, two-player, zero-sum game. We use two types of game tree search algorithms to solve this problem: minimax search tree and Monte-Carlo search tree. Both search trees can yield the optimal policy but may require possibly exponential computational time and space. We first propose three pruning techniques to reduce the computational time while preserving optimality guarantees. When the agent and the opponent are located far from each other initially, we present a variable resolution technique with longer planning horizon to further reduce computational time. Simulation results show the effectiveness of the proposed strategies in terms of computational time.

Deploying a team of robots that can carefully coordinate their actions can make the entire system robust to individual failures. In this report, we review recent algorithmic development in making multi-robot systems robust to environmental uncertainties, failures, and adversarial attacks. We find the following three trends in the recent research in the area of multi-robot coordination: (1) resilient coordination to either withstand failures and/or attack or recover from failures/attacks; (2) risk-aware coordination to manage the trade-off risk and reward, where the risk stems due to environmental uncertainty; (3) Graph Neural Networks based coordination to learn decentralized multi-robot coordination policies. These algorithms have been applied to tasks such as formation control, task assignment and scheduling, search and planning, and informative data collection. In order for multi-robot systems to become practical, we need coordination algorithms that can scale to large teams of robots dealing with dynamically changing, failure-prone, contested, and uncertain environments. There has been significant recent research on multi-robot coordination that has contributed resilient and risk-aware algorithms to deal with these issues and reduce the gap between theory and practice. Learning-based approaches have been seen to be promising, especially since they can learn who, when, and how to communicate for effective coordination. However, these algorithms have also been shown to be vulnerable to adversarial attacks, and as such developing learning-based coordination strategies that are resilient to such attacks and robust to uncertainties is an important open area of research.

The ability to apply controlled forces to individual molecules has been revolutionary in shaping our understanding of biophysics in areas as diverse as dynamic bond strength, biological motor operation, and DNA replication. However, the methodology to perform single-molecule experiments remains relatively inaccessible because of cost and complexity. In 2010, we introduced the centrifuge force microscope (CFM) as a platform for accessible and high-throughput single-molecule experimentation. The CFM consists of a rotating microscope with which prescribed centrifugal forces can be applied to microsphere-tethered biomolecules. In this work, we develop and demonstrate a next-generation Wi-Fi CFM that offers unprecedented ease of use and flexibility in design. The modular CFM unit fits within a standard benchtop centrifuge and connects by Wi-Fi to an external computer for live control and streaming at near gigabit speeds. The use of commercial wireless hardware allows for flexibility in programming and provides a streamlined upgrade path as Wi-Fi technology advances. To facilitate ease of use, detailed build and setup instructions, as well as LabVIEW-based control software and MATLAB-based analysis software, are provided. We demonstrate the instrument’s performance by analysis of force-dependent dissociation of short DNA duplexes of 7, 8, and 9 bp. We showcase the sensitivity of the approach by resolving distinct dissociation kinetic rates for a 7 bp duplex in which one G-C basepair is mutated to an A-T basepair.

The ability to apply controlled forces to individual molecules has been revolutionary in shaping our understanding of biophysics in areas as diverse as dynamic bond strength, biological motor operation, and DNA replication. However, the methodology to perform single-molecule experiments has been and remains relatively inaccessible due to cost and complexity. In 2010, we introduced the Centrifuge Force Microscope (CFM) as a new platform for accessible and high-throughput single-molecule experimentation. The CFM consists of a rotating microscope where prescribed centrifugal forces can be applied to microsphere-tethered biomolecules. In this work, we develop and demonstrate a next-generation Wi-Fi CFM that offers unprecedented ease of use and flexibility in design. The modular CFM unit fits within a standard benchtop centrifuge and connects by Wi-Fi to a external computer for live control and streaming at near gigabit speeds. The use of commercial wireless hardware allows for flexibility in programming and provides a streamlined upgrade path as Wi-Fi technology improves. To facilitate ease of use, detailed build and setup instructions are provided, as well as LabVIEW™ based control software and MATLAB® based analysis software. We demonstrate the analysis of force-dependent dissociation of short DNA duplexes of 7, 8, and 9 bp using the instrument. We showcase the sensitivity of the approach by resolving distinct dissociation kinetic rates for a 7 bp duplex where one G-C base pair is mutated to an A-T base pair.SignificanceThe ability to apply mechanical forces to individual molecules has provided unprecedented insight into many areas of biology. Centrifugal force provides a way to increase the throughput and to decrease the cost and complexity of single-molecule experiments compared to other approaches. In this work, we develop and demonstrate a new user-friendly Centrifuge Force Microscope (CFM) that enables live-streaming of high-throughput single-molecule experiments in a benchtop centrifuge. We achieved near gigabit bandwidth with standard Wi-Fi components, and we provide detailed design instructions and software to facilitate use by other labs. We demonstrate the instrument for sensitive kinetic measurements that are capable of resolving the difference between two DNA duplexes that differ by a single G-C to A-T substitution.

In this paper, we design algorithms to protect swarm-robotics applications against sensor denial-of-service (DoS) attacks on robots. We focus on applications requiring the robots to jointly select actions, e.g., which trajectory to follow, among a set of available ones. Such applications are central in large-scale robotic applications, such as multi-robot motion planning for target tracking. But the current attack-robust algorithms are centralized. In this paper, we propose a general-purpose distributed algorithm towards robust optimization at scale, with local communications only. We name it Distributed Robust Maximization (DRM). DRM proposes a divide-and-conquer approach that distributively partitions the problem among cliques of robots. Then, the cliques optimize in parallel, independently of each other. We prove DRM achieves a close-to-optimal performance. We demonstrate DRM's performance in both Gazebo and MATLAB simulations, in scenarios of active target tracking with swarms of robots. In the simulations, DRM achieves computational speed-ups, being 1-2 orders faster than the centralized algorithms; yet, it nearly matches the tracking performance of the centralized counterparts. Since, DRM overestimates the number of attacks in each clique, in this paper we also introduce an Improved Distributed Robust Maximization (IDRM) algorithm. IDRM infers the number of attacks in each clique less conservatively than DRM by leveraging 3-hop neighboring communications. We verify IDRM improves DRM's performance in simulations.

Viral detection is critical for controlling disease spread and progression. Recent emerging viral threats including Zika, Ebola, and the current COVID-19 outbreak highlight the cost and difficulty in responding rapidly. To address these challenges, we develop a platform for low-cost and rapid detection of viral RNA with DNA nanoswitches designed to mechanically reconfigure in response to specific viruses. Using Zika virus as a model system, we show non-enzymatic detection of viral RNA to the attomole level, with selective and multiplexed detection between related viruses and viral strains. For clinical-level sensitivity in biological fluids, we paired the assay with a sample preparation step using either RNA extraction or isothermal pre-amplification. Our assay can be performed with minimal or no lab infrastructure, and is readily adaptable to detect other viruses. We demonstrate the adaptability of our method by quickly developing and testing DNA nanoswitches for detecting a fragment of SARS-CoV-2 RNA in human saliva. Given this versatility, we expect that further development and field implementation will improve our ability to detect emergent viral threats and ultimately limit their impact.

The multiple-path orienteering problem asks for paths for a team of robots that maximize the total reward collected while satisfying budget constraints on the path length. This problem models many multi-robot routing tasks such as exploring unknown environments and information gathering for environmental monitoring. In this paper, we focus on how to make the robot team robust to failures when operating in adversarial environments. We introduce the Robust Multiple-path Orienteering Problem (RMOP) where we seek worst-case guarantees against an adversary that is capable of attacking at most $\alpha$ robots. We consider two versions of this problem: RMOP offline and RMOP online. In the offline version, there is no communication or replanning when robots execute their plans and our main contribution is a general approximation scheme with a bounded approximation guarantee that depends on $\alpha$ and the approximation factor for single robot orienteering. In particular, we show that the algorithm yields a (i) constant-factor approximation when the cost function is modular; (ii) $\log$ factor approximation when the cost function is submodular; and (iii) constant-factor approximation when the cost function is submodular but the robots are allowed to exceed their path budgets by a bounded amount. In the online version, RMOP is modeled as a two-player sequential game and solved adaptively in a receding horizon fashion based on Monte Carlo Tree Search (MCTS). In addition to theoretical analysis, we perform simulation studies for ocean monitoring and tunnel information-gathering applications to demonstrate the efficacy of our approach., Comment: submitted to TRO

We study the problem of incorporating risk while making combinatorial decisions under uncertainty. We formulate a discrete submodular maximization problem for selecting a set using Conditional-Value-at-Risk (CVaR), a risk metric commonly used in financial analysis. While CVaR has recently been used in optimization of linear cost functions in robotics, we take the first stages towards extending this to discrete submodular optimization and provide several positive results. Specifically, we propose the Sequential Greedy Algorithm that provides an approximation guarantee on finding the maxima of CVaR cost function under a matroidal constraint. The approximation guarantee shows that the solution produced by our algorithm is within a constant factor of the optimal and an additive term that depends on the optimal. Our analysis uses the curvature of the submodular set function, and proves that the algorithm runs in polynomial time. This formulates a number of combinatorial optimization problems that appear in robotics. We use two such problems, vehicle assignment under uncertainty for mobility-on-demand and sensor selection with failures for environmental monitoring, as case studies to demonstrate the efficacy of our formulation.

Submodular maximization has been widely used in many multi-robot task planning problems including information gathering, exploration, and target tracking. However, the interplay between submodular maximization and communication is rarely explored in the multi-robot setting. In many cases, maximizing the submodular objective may drive the robots in a way so as to disconnect the communication network. Driven by such observations, in this paper, we consider the problem of maximizing submodular function with connectivity constraints. Specifically, we propose a problem called Communication-aware Submodular Maximization (CSM), in which communication maintenance and submodular maximization are jointly considered in the decision-making process. One heuristic algorithm that consists of two stages, i.e. \textit{topology generation} and \textit{deviation minimization} is proposed. We validate the formulation and algorithm through numerical simulation. We find that our algorithm on average suffers only slightly performance decrease compared to the pure greedy strategy., Comment: accepted to ICRA2021

Due to the fast learning speed, simplicity of implementation and minimal human intervention, extreme learning machine has received considerable attentions recently, mostly from the machine learning community. Generally, extreme learning machine and its various variants focus on classification and regression problems. Its potential application in analyzing censored time-to-event data is yet to be verified. In this study, we present an extreme learning machine ensemble to model right-censored survival data by combining the Buckley-James transformation and the random forest framework. According to experimental and statistical analysis results, we show that the proposed model outperforms popular survival models such as random survival forest, Cox proportional hazard models on well-known low-dimensional and high-dimensional benchmark datasets in terms of both prediction accuracy and time efficiency.

We study a distributed coordination problem in which a group of robots collaborates to maximize a submodular objective function. We consider an adversarial scenario where a subset of the robots can be attacked and cannot contribute to the objective function. We do not know which robots will be attacked, but know an upper bound of how many can be attacked. We study the distributed version of the problem, where each robot must choose its own actions only by communicating with its neighbors. We present a distributed resilient submodular maximization algorithm that guarantees performance within a constant factor of the optimal. Our analysis uses the curvature of submodular set functions. We show that the algorithm is scalable, runs in polynomial time, and is faster than its centralized version. We demonstrate the efficacy of our algorithm through simulations of a multi-robot target tracking scenario.

We introduce and study the problem of planning a trajectory for an agent to carry out a reconnaissance mission while avoiding being detected by an adversarial guard. This introduces a multi-objective version of classical visibility-based target search and pursuit-evasion problem. In our formulation, the agent receives a positive reward for increasing its visibility (by exploring new regions) and a negative penalty every time it is detected by the guard. The objective is to find a finite-horizon path for the agent that balances the trade off between maximizing visibility and minimizing detectability.We model this problem as a discrete, sequential, two-player, zero-sum game. We use two types of game tree search algorithms to solve this problem: minimax search tree and Monte-Carlo search tree. Both search trees can yield the optimal policy but may require possibly exponential computational time and space. We propose several pruning techniques to reduce the computational cost while still preserving optimality guarantees. Simulation results show that the proposed strategy prunes approximately three orders of magnitude nodes as compared to the brute-force strategy. We also find that the Monte-Carlo search tree saves approximately one order of computational time as compared to the minimax search tree.

Cloud computing enables customers with limited computational resources to outsource their huge computation workloads to the cloud with massive computational power. However, in order to utilize this computing paradigm, it presents various challenges that need to be addressed, especially security. As eigen-decomposition (ED) and singular value decomposition (SVD) of a matrix are widely applied in engineering tasks, we are motivated to design secure, correct, and efficient protocols for outsourcing the ED and SVD of a matrix to a malicious cloud in this paper. In order to achieve security, we employ efficient privacy-preserving transformations to protect both the input and output privacy. In order to check the correctness of the result returned from the cloud, an efficient verification algorithm is employed. A computational complexity analysis shows that our protocols are highly efficient. We also introduce an outsourcing principle component analysis as an application of our two proposed protocols.

Significant progress in DNA nanotechnology has accelerated the development of molecular machines with functions like macroscale machines. However, the mobility of DNA self-assembled nanorobots is still dramatically limited due to challenges with designing and controlling nanoscale systems with many degrees of freedom. Here, an origami-inspired method to design transformable DNA nanomachines is presented. This approach integrates stiff panels formed by bundles of double-stranded DNA connected with foldable creases formed by single-stranded DNA. To demonstrate the method, a DNA version of the paper origami mechanism called a waterbomb base (WBB) consisting of six panels connected by six joints is constructed. This nanoscale WBB can follow four distinct motion paths to transform between five distinct configurations including a flat square, two triangles, a rectangle, and a fully compacted trapezoidal shape. To achieve this, the sequence specificity of DNA base-pairing is leveraged for the selective actuation of joints and the ion-sensitivity of base-stacking interactions is employed for the flattening of joints. In addition, higher-order assembly of DNA WBBs into reconfigurable arrays is achieved. This work establishes a foundation for origami-inspired design for next generation synthetic molecular robots and reconfigurable nanomaterials enabling more complex and controllable motion.

Communication modernization is an important symbol of human societies as we enter into the information age. An importantproblem encountered in modern communication is interference. With the development of communication, the establishment of variouscommunication networks, making the limited frequency of resources more crowded, intertwined to the serious. How to ensure that thecommunication is effective, accurate and rapid in the harsh environment conditions in front of today's communications researchers is adifficult problem. As the mobile communication system must use wireless transmission technology in order to achieve the exchange ofinformation in the mobile, wireless transmission is vulnerable to a variety of other radio waves interference, a large number of interferencewill greatly affect the quality of network communications and system capacity. Anti-jamming technology with strong resistance, strongtechnical, difficult, practical and reliable high characteristics, in today's increasingly harsh electromagnetic environment, anti-jammingtechnology is particularly compelling. It can greatly improve the effectiveness and reliability of the communication system in thecommunication system. Today, anti-jamming technology has become the mainstream of radio communications technology, it is the mostclosely linked with the field of communications. This paper mainly discusses the interference problem of GSM system in mobilecommunication anti-jamming technology, including the interference type in GSM, the anti-interference ability of GSM itself and how tooptimize the network to achieve the purpose of anti-jamming, so as to meet the requirements of users System capacity while maintaininggood call quality.

We study the problem of designing spoofing signals to corrupt and mislead the output of a Kalman filter. Unlike existing works that focus on detection and filtering algorithms for the observer, we study the problem from the attacker's point-of-view. In our model, the attacker can corrupt the measurements by adding spoofing signals. The attacker seeks to create a separation between the estimate of the Kalman filter with and without spoofing signals. We present a number of results on how to generate such spoofing signals, while minimizing the signal magnitude. The resulting algorithms are evaluated through simulations along with theoretical proofs.

The problem of target tracking with multiple robots consists of actively planning the motion of the robots to track the targets. A major challenge for practical deployments is to make the robots resilient to failures. In particular, robots may be attacked in adversarial scenarios, or their sensors may fail or get occluded. In this letter, we introduce planning algorithms for multi-target tracking that are resilient to such failures. In general, resilient target tracking is computationally hard. Contrary to the case where there are no failures, no scalable approximation algorithms are known for resilient target tracking when the targets are indistinguishable, or unknown in number, or with unknown motion model. In this letter, we provide the first such algorithm, which also has the following properties: First, it achieves maximal resiliency, since the algorithm is valid for any number of failures. Second, it is scalable, as our algorithm terminates with the same running time as state-of-the-art algorithms for (non-resilient) target tracking. Third, it provides provable approximation bounds on the tracking performance, since our algorithm guarantees a solution that is guaranteed to be close to the optimal. We quantify our algorithm's approximation performance using a novel notion of curvature for monotone set functions subject to matroid constraints. Finally, we demonstrate the efficacy of our algorithm through MATLAB and Gazebo simulations and a sensitivity analysis; we focus on scenarios that involve a known number of distinguishable targets.

This study proposes a distributed model predictive control (MPC) strategy to achieve consensus of sampled-data multi-agent systems with double-integrator dynamics. On the basis of the error of state between each agent and the centre of its subsystem, a novel distributed MPC strategy (Algorithm 1) is obtained with the exchange of current states only. Then, a reverse iterative algorithm (Algorithm 2) is specially designed for the receding horizon optimisation of sampled-data double-integrator dynamics. Illustrative examples are finally displayed to verify the effectiveness and advantage of the distributed MPC consensus strategy and the impact of sampling period on consensus.

Cloud computing provides service for resource-constrained customers to perform large-scale scientific computation. However, it also brings some new challenges, which have to be considered in designing outsourcing protocols. In recent years, a few outsourcing protocols have been proposed for different kinds of problems. Quadratic programming (QP) is a class of mathematical optimization problem, and solving a large-scale QP problem requires a large amount of computation. Thus, there is a great need for customer to outsource large-scale QP problem to cloud. In this paper, we design a secure, verifiable, and efficient outsourcing protocol for QP problem. For security consideration, we encrypt the matrices and vectors contained in the QP problem in an efficient way. After cloud computing, we decrypt the result to get the ultimate solution. To ensure correctness, we verify the result returned by the cloud through Karush-Kuhn-Tucker conditions that are the necessary and sufficient conditions for the optimal solution. We also present complexity analysis and numerical simulations to illustrate the efficiency of our outsourcing protocol.

We study the problem of reducing the amount of communication in a distributed target tracking problem. We focus on the scenario where a team of robots are allowed to move on the boundary of the environment. Their goal is to seek a formation so as to best track a target moving in the interior of the environment. The robots are capable of measuring distances to the target. Decentralized control strategies have been proposed in the past that guarantee that the robots asymptotically converge to the optimal formation. However, existing methods require that the robots exchange information with their neighbors at all time steps. Instead, we focus on reducing the amount of communication among robots. We propose a self-triggered communication strategy that decides when a particular robot should seek up-to-date information from its neighbors and when it is safe to operate with possibly outdated information from the neighbor. We prove that this strategy converges to an optimal formation. We compare the two approaches (constant communication and self-triggered communication) through simulations of tracking stationary and mobile targets.

We study the problem of reducing the amount of communication in decentralized target tracking. We focus on the scenario where a team of robots are allowed to move on the boundary of the environment. Their goal is to seek a formation so as to best track a target moving in the interior of the environment. The robots are capable of measuring distances to the target. Decentralized control strategies have been proposed in the past that guarantee that the robots asymptotically converge to the optimal formation. However, existing methods require that the robots exchange information with their neighbors at all time steps. Instead, we focus on decentralized strategies to reduce the amount of communication among robots. We propose a self-triggered communication strategy that decides when a particular robot should seek up-to-date information from its neighbors and when it is safe to operate with possibly outdated information. We prove that this strategy converges asymptotically to a desired formation when the target is stationary. For the case of a mobile target, we propose an extension whereby each robot decides its optimal partner to share its measurements with using observability as a criterion. We evaluate all the approaches (constant communication and self-triggered communication with centralized and decentralized sensor fusion) through simulations., Comment: Our paper has been accepted as a regular paper and is going to be published in IEEE Transactions on Automation Science and Engineering. We update the final version of our paper. The main change from previous arXiv version is the discussion about the limited communication and sensing range(Section V-B, page 6) and more details about the calculation of the omega_max in the appendix

We study two sensor assignment problems for multi-target tracking with the goal of improving the observability of the underlying estimator. We consider various measures of the observability matrix as the assignment value function. We first study the general version where the sensors must form teams to track individual targets. If the value function is monotonically increasing and submodular then a greedy algorithm yields a 1/2-approximation. We then study a restricted version where exactly two sensors must be assigned to each target. We present a 1/3-approximation algorithm for this problem which holds for arbitrary value functions (not necessarily submodular or monotone). In addition to approximation algorithms, we also present various properties of observability measures. We show that the inverse of the condition number of the observability matrix is neither monotone nor submodular, but present other measures which are. Specifically, we show that the trace and rank of the symmetric observability matrix are monotone and submodular and the log determinant of the symmetric observability matrix is monotone and submodular when the matrix is non-singular. If the target's motion model is not known, the inverse cannot be computed exactly. Instead, we present a lower bound for distance sensors. In addition to theoretical results, we evaluate our results empirically through simulations., Comment: Second version submission. Accidental submission as new article. Please see arXiv:1706.00871

Prediction of seminal quality with statistical learning tools is an emerging methodology in decision support systems in biomedical engineering and is very useful in early diagnosis of seminal patients and selection of semen donors candidates. However, as is common in medical diagnosis, seminal quality prediction faces the class imbalance problem. In this paper, we propose a novel supervised ensemble learning approach, namely Clustering-Based Decision Forests, to tackle unbalanced class learning problem in seminal quality prediction. Experiment results on real fertility diagnosis dataset have shown that Clustering-Based Decision Forests outperforms decision tree, Support Vector Machines, random forests, multilayer perceptron neural networks and logistic regression by a noticeable margin. Clustering-Based Decision Forests can also be used to evaluate variables’ importance and the top five important factors that may affect semen concentration obtained in this study are age, serious trauma, sitting time, the season when the semen sample is produced, and high fevers in the last year. The findings could be helpful in explaining seminal concentration problems in infertile males or pre-screening semen donor candidates.

Prediction capability of a classifier usually improves when it is built from an extended variable space by adding new variables from randomly combination of two or more original variables. However, its usefulness in survival analysis of censored time-to-event data is yet to be verified. In this research, we investigate the plausibility of space extension technique, originally proposed for classification purpose, to survival analysis. By combing random subspace, bagging and extended space techniques, we develop a random survival forest with space extensions algorithm. According to statistical analysis results, we show that the proposed model outperforms or at least comparable to popular survival models such as random survival forest, rotation survival forest, Cox proportional hazard and boosting survival models on well-known benchmark datasets.

Multimedia transmission over the network is susceptible to various runtime impairments such as process failure, network congestion or link error. Existing work usually determines such quality of service (QoS) violations through condition–action rules, which trigger corresponding actions once pre-described conditions are satisfied. However, the results of such rigid rules are often not satisfactory in practice in that there has been little serious study with respect to the relationship between the root cause of a QoS violation and the observed violation phenomenon. In this paper, we introduce a statistical approach to the analysis of QoS violations. We propose and validate through experiments that: (1) a type of QoS violation will present consistent symptoms in terms of the observed application performance and end-to-end traffic pattern. Such a violation can be recognized once the similar symptoms repeat during a QoS session. (2) QoS violations of different nature (e.g. caused by shortage of different resources) will present diverse symptoms. Using a set of end-to-end statistics, we are able to describe and differentiate between QoS violations. We propose a fast orthonormal algorithm for real-time training/classification of QoS violations and prove that this algorithm is universal approximation. We also extend the scope of hidden neurons from kernel functions to additive functions for higher classification accuracy. Our experiments validate our ideas and show the effectiveness of our approach in terms of training overhead and classification accuracy. Copyright © 2011 John Wiley & Sons, Ltd.

This paper presents a distributed model predictive control(DMPC) strategy to achieve cooperative tasks like consensus and synchronization of linear systems with coupled constraints. Based on the error of state between each subsystem and the center of its cooperative set, a novel DMPC algorithm is designed with its robust feasibility guaranteed by permitting only one subsystem to optimize at each time instant, while renew other subsystems' plans. Illustrative examples are finally conducted to verify the effectiveness of the DMPC strategy.

Recently, survival ensembles have found more and more applications in biological and medical research when censored time-to-event data are often confronted. In this research, we investigate the plausibility of extending rotation forest, originally proposed for classification purpose, to survival analysis. Supported by the proper statistical analysis, we show that rotation survival forests are able to outperform the state-of-art survival ensembles on right censored data. We also provide a C-index based variable importance measure for evaluating covariates in censored survival data.

Recently, various ensemble learning methods with different base classifiers have been proposed for credit scoring problems. However, for various reasons, there has been little research using logistic regression as the base classifier. In this paper, given large unbalanced data, we consider the plausibility of ensemble learning using regularized logistic regression as the base classifier to deal with credit scoring problems. In this research, the data is first balanced and diversified by clustering and bagging algorithms. Then we apply a Lasso-logistic regression learning ensemble to evaluate the credit risks. We show that the proposed algorithm outperforms popular credit scoring models such as decision tree, Lasso-logistic regression and random forests in terms of AUC and F-measure. We also provide two importance measures for the proposed model to identify important variables in the data.

To surface the Deep Web, one crucial task is to predict whether a given web page has a search interface (searchable HyperText Markup Language (HTML) form) or not. Previous studies have focused on supervised classification with labeled examples. However, labeled data are scarce, hard to get and requires tediousmanual work, while unlabeled HTML forms are abundant and easy to obtain. In this research, we consider the plausibility of using both labeled and unlabeled data to train better models to identify search interfaces more effectively. We present a semi-supervised co-training ensemble learning approach using both neural networks and decision trees to deal with the search interface identification problem. We show that the proposed model outperforms previous methods using only labeled data. We also show that adding unlabeled data improves the effectiveness of the proposed model.

In this paper, we propose an improved random forest algorithm which allocates weights to decision trees in the forest during tree aggregation for prediction and their weights are easily calculated based on out-of-bag errors in training. We compare the performance of our proposed algorithm and the original one on loan default prediction datasets . We also use these two algorithms to create two kinds of balanced random forests to deal with imbalanced data problem. Experiments results show that our proposed algorithm beat s the original r andom f orest in terms of both balanced and overall accuracy metric s . Experiments also show that parallel r andom f orests can greatly improve r andom forests’ efficiency during the learning process. DOI: http://dx.doi.org/10.11591/telkomnika.v10i6.1323

Multi-hop relaying is an important concept in future generation wireless networks. It can address the inherent problems of limited capacity and coverage in cellular networks. However, most existing multi-hop relaying architectures waste energy. In this paper, our goal is to find a power-effective multi-hop relaying architecture, since the problem of coverage extension may occur in the given target service area where users' traffic density is relatively low or moderate, this paper introduces several popular topologies and the corresponding coverage extension methods for each case. Further, sets up the optimized mathematical model and obtain the simulation results which show that Type-D is the best topology which consumes less power than Type-A, Type-B and Type-C.

In this paper, we introduce a novel end-to-end approach to QoS management with respect to the diagnosis of QoS violations. We first use a set of end-to-end flow traffic statistics to describe a QoS violation. Subsequently, neural network techniques are engaged to identify and differentiate QoS violations through classification of the collected statistics. Through experiments, we find that our scheme outperforms traditional rule-based methods which require clear margins of QoS parameters in asserting a QoS violation.

In this paper, we propose a novel cross-domain QoS framework for coordinated end-to-end QoS adaptation. The framework treats each autonomous QoS service on the end-to-end path as a QoS component and models its capabilities and behavior with respect to the requirements of end-to-end QoS coordination. Based on such information, the end-to-end QoS framework runtime copes with a QoS violation by selecting a suitable QoS adaptation solution among all the end-to-end candidate choices. Preliminary simulation results demonstrate the feasibility of our framework for distributed end-to-end QoS coordination.

Quality of Service is commonly specified through either a set of pre-defined APIs or a specification language. The former approach has inherent limitations that make QoS specification heuristic, while the latter largely remains at syntax level, which is strong in representation formats, but weak in the semantic interpretation of QoS specification. To overcome these limitations, we propose here a semantic QoS specification (SQS) framework, which (1) provides a uniform framework for QoS specification, (2) offers tailored specification support for different application domains, and (3) facilitates machine processing of QoS specification. We adopt an ontology-based approach in the design of SQS and introduce two generic models for a more flexibile, accurate and expressive QoS specification. Moreover, the ontology modeling principle of SQS is sufficiently general to apply to the modeling of other QoS systems, which facilitates information sharing and negotiation among QoS entities along the end-to-end path. Our experimental results show that SQS have exhibited the above advantages for QoS specification with acceptable overhead in term of processing latency.

In this paper, we propose a formal knowledge modeling methodology based on ontology to assist in the logic reasoning of distributed QoS information which is commonly not available in traditional QoS modeling and support. As part of our project toward a semantic-based QoS control and management framework, the knowledge modeling scheme presented in this paper enables an explicit representation of various QoS concepts and their relations, thus provides a foundation for understanding and decision-making of QoS during setup and runtime. The technical feasibility of our knowledge-based QoS scheme has been demonstrated in our prototype, which yields the aforementioned benefits with acceptable processing overhead on PCs and laptops.

Future distributed applications are expected to be deployed in an environment that is more dynamic and heterogeneous than ever before. The environment features are difficult to predict beforehand due to their variations. To cope with such variations, it is important that the protocol stack is adaptable to changing requirements. This paper proposes a component-based Dynamical Protocol Framework (DPF) for multimedia applications. DPF dynamically builds an adaptive protocol stack by automatic discovery of protocol components and runtime configuration/reconfiguration of the protocol stack. Performance measurement shows that compared with directly using interfaces provided by JMF, using DPF to set up a dynamic protocol stack incurs slight overhead in unicast, and a similar condition is observed in multicast. This is the trade-off for the abilities to configure and reconfigure protocol stack during stream set-up and run time.

Recognizing the need to provide end-to-end QoS for performance sensitive applications, and the need to manage such QoS provisions in the presence of diverse application QoS requirements as well as heterogeneous network environment, we propose an adaptive end-to-end QoS management architecture QMan, to address such issues. QMan, as one of the key components of OCTOPUS middleware, makes no presumption of the underneath QoS provisions for end-to-end QoS management, but can take advantage of such provisions if there are. QMan allows applications to exchange QoS requirements and negotiate for a suitable protocol stack configuration with the assistance from dynamic protocol framework (DPF). During data transmission, QMan monitors QoS violations and performs QoS adaptations according to the rules defined through a combination of three possible conduits in an end-to-end fashion: network level, host level and application level. These rules are pre-determined according to application requirements or network policies and can be re-configured for customization of QoS adaptations during run-time. Performance evaluations of our implementation demonstrate the feasibility of QMan's design with satisfactory performance.