This paper discusses a fully decentralized online federated learning setting with long-term constraints. The fully decentralized setting removes communication and computational bottlenecks associated with a central server communicating with a large number of clients. Also, online learning is introduced to the federated learning setting to capture situations with time-varying data distribution. Practical federated learning settings are imposed with long-term constraints such as energy constraints, money cost constraints, time constraints etc. The clients are not obligated to satisfy any per round constraint, but they must satisfy these long-term constraints. To provide privacy to the shared local model updates of the clients, local differential privacy is introduced. An online mirror descent-based algorithm is proposed and its regret bound is obtained. The regret bound is compared with the regret bound of a differentially-private version of online gradient descent algorithm proposed for federated learning.

Pressure-sensitive smart floors deployed within homes can give great insight to the health and activity level of individuals through gait and location information. Due to the ever-changing dynamic nature of household deployments involving furniture movement, floor tile shifts, and sensor drift, challenges arise in ensuring the constant reliability of floor sensor readings over time. This paper presents a procedure to automatically calibrate a smart floor’s force sensors without specialized physical effort. The calibration algorithm automatically filters out non-human static weight while retaining weight generated by human activity. This technique is designed to correctly translate sensor values to weight units even when direct access to the force sensors is not available and when a shared tile floor sits above the sensor grid. These calibrated sensor values can then feed machine learning techniques used to extract individual contact points generated by a person’s walking cycle. Using known human weights but no knowledge of the human’s location or walking trajectory, this calibration technique resulted in small percentage differences of -7.8%, -4.8%, and -1.6% for the mean, median, and mode of calibrated smart floor walking sequences, respectively.

This is the first study of crowdsourcing Pareto-optimal object finding over partial orders and by pairwise comparisons, which has applications in public opinion collection, group decision making, and information exploration. Departing from prior studies on crowdsourcing skyline and ranking queries, it considers the case where objects do not have explicit attributes and preference relations on objects are strict partial orders. The partial orders are derived by aggregating crowdsourcers' responses to pairwise comparison questions. The goal is to find all Pareto-optimal objects by the fewest possible questions. It employs an iterative question-selection framework. Guided by the principle of eagerly identifying non-Pareto optimal objects, the framework only chooses candidate questions which must satisfy three conditions. This design is both sufficient and efficient, as it is proven to find a short terminal question sequence. The framework is further steered by two ideas---macro-ordering and micro-ordering. By different micro-ordering heuristics, the framework is instantiated into several algorithms with varying power in pruning questions. Experiment results using both real crowdsourcing marketplace and simulations exhibited not only orders of magnitude reductions in questions when compared with a brute-force approach, but also close-to-optimal performance from the most efficient instantiation.

On behalf of the entire Organization Committee, it is our pleasure to welcome you to the Twelth IEEE International Conference on Pervasive Computing and Communications (PerCom 2014), sponsored by IEEE Computer Society, the University of Texas at Arlington (UTA), Budapest University of Technology and Economics (BME) and the Scientific Associations for Infocommunications Hungary (HTE). The Proceedings confirm the high quality tradition of PerCom that makes this Conference one of the premier scholarly venues in the world for pervasive computing and communications. This edition of the Conference takes place in Budapest, the capital of Hungary, a lovely city in the heart of Europe.

In this paper, we address the problem of channel assignment considering partially overlapping channels (POCs) for interference avoidance in 802.11-based wireless mesh networks. A novel interference model is proposed which provides a systematic approach of measuring the interference caused by links operating on POCs by taking into account both the adjacent channel interference and the corresponding physical distance between mesh nodes. Based on this model, we design a centralized and a distributed interference-aware channel assignment algorithm called i-POCA which enables the use of smart ants for assigning orthogonal and non-orthogonal channels to radios in order to minimize total network interference. We evaluate our algorithms through extensive simulations and demonstrate that our proposed algorithms improve network throughput by efficient utilization of the available spectrum.

Communication is an important element of multiagent systems (MAS). In fully decentralized systems it is needed to allow the agents to coordinate their actions to achieve certain goals. When the agents have no means to coordinate their actions, they generally choose actions that minimize their chance of losses. If the agents were allowed to coordinate, on the other hand, they can choose actions that allow them to get higher rewards instead. In game theory, these concepts are known as risk dominance and payoff dominance. In this paper, we model communication between agents in stochastic games as an extensive form game in which the agents can numerically evaluate the benefit of communicating a piece of information to the other agents. This allows the agent to answer two important questions about the communication process, namely when should an agent communicate, and what should an agent communicate. We also present a more stable way to select between Nash equilibria in multiagent reinforcement learning using NashQ which is used in the calculation of the values for the communication game.

The objective of inverse reinforcement learning (IRL) is to learn an agent's reward function based on either the agent's policies or the observations of the policy. In this paper we address the issue of using inverse reinforcement learning to learn the reward function in a multi agent setting, where the agents can either cooperate or be strictly non-cooperative. The case of cooperataing agents is a subcase of the non-cooperative setting, where the agents collectively try to maximize a common reward function, instead of maximizing their individual reward functions. Here we present an IRL algorithm that considers the case where the policies of the agents are known. We use the framework that was described by Ng and Russell [2001] and extend it for a Multiagent setting. We assume that the agents are rational and follow an optimal policy in the sense of the Nash Equilibrium. These assumptions are very common in Multiagent systems. We show that in the case of known policies we can reduce the Multiagent problem to a distributed solution where the reward function for each agent can be solved independently using a very similar formulation as for the single agent case.

Load balancing in wireless mesh networks (WMNs) is one of the major design goals for any routing protocol. In this paper, we propose AntMesh, an efficient data forwarding scheme, specifically designed for load balancing in multi-radio infrastructure WMNs. As the foundation of AntMesh, we use Ant Colony Optimization (ACO) in which artificial ants (agents) perform the routing and data forwarding to stochastically solve a dynamic network optimization problem. We extend the ACO framework by implementing two estimation modules, i.e., a link and a path estimation module to effectively utilize the space/channel diversity typically common in multi-radio WMNs. The link estimation module measures the cost of a node's local links in terms of the packet delay taking into account the queuing delay of a node to realistically capture load. The path estimation module captures inter/intra flow interference by selecting a reduced interference path with increased channel diversity thus resulting in improved load balancing. Simulation studies comparing AntMesh to competing approaches are provided. We demonstrate that under high loads, AntMesh provides increased throughput and decreased end-to-end delay compared to these other approaches.

In this paper, we address the problem of packet routing for interference avoidance in infrastructure mesh networks. We propose AMIRA, an interference-aware routing protocol designed to improve load balancing by avoiding inter and intra flow interference in a typical mesh backbone network. AMIRA is based on the framework of ant colony optimization (ACO) which is a meta-heuristic approach for stochastically solving a problem together with our local heuristic technique to avoid interference within and among packet flows. In AMIRA, each node uses MAC level information to measure link qualities which helps in selecting reduced interference paths thus resulting in improved load balancing in addition to the auto load balancing feature of the ACO framework. Our simulation results demonstrate that under congestion, AMIRA gives increased throughput and low end-to-end delay when compared to other existing routing protocols because of its interference aware technique and stochastic data forwarding nature.

In this paper, we are proposing a scheme called RECHOKe (REpeatedly CHOose and keep for malicious flows, REpeatedly CHOose and Kill for non-malicious flows) to be used for detecting, controlling and punishing of malicious flows in IP networks. It is an extension of xCHOKe, CHOKe and RED-PD schemes, combining both CHOKe hit and RED drop/mark histories, to detect, control and punish these flows more accurately while providing better protection to non-malicious flows. However, unlike xCHOKe and CHOKe, RECHOKe does not drop packets during CHOKe hits; thereby eliminating the complexity of dropping or marking randomly selected packets already queued and the unreliability of CHOKe hits. We analyze xCHOKe and RECHOKe in detail using ns-2 and show that RECHOKe performs better than RED, CHOKe and xCHOKe which are limited in what they can achieve as malicious flows get much more than their fair share and non-malicious flows get mistakenly penalized.

An interferometric ranging technique has been recently proposed as a possible way to localize ad hoc and sensor networks. Compared to the more common techniques such as received signal strength, time of arrival, and angle of arrival ranging, interferometric ranging has the advantage that the measurement could be highly precise. However, localization using interferometric ranging is difficult as it requires a large number of measurement readings. In this paper, we provide a formal proof of this difficulty in terms of algorithmic complexity. Furthermore, we propose an iterative algorithm that calculates node locations from a set of seeding anchors, gradually building a more global localization solution. Compared to previous localization algorithms, which treat localization as a global optimization problem, the iterative algorithm is a distributed algorithm that is simple to implement in larger networks. More importantly, the iterative algorithm allows us to study the error propagation behavior of localization using interferometric ranging. Using simulations, we validate the performance of the iterative algorithm in terms of localization error and coverage.

Localization in sensor networks is the process of obtaining geographical location information for all deployed sensors. Localizing each sensor node is becoming increasingly important as more and more algorithms and protocols in the disciplines of routing, energy management, and security have been proposed that rely on (semi-) accurate location information. In this paper, we study a particular approach to sensor node localization that uses sequentially deployed beacons to localize sensors. Our model assumes that the localization information can be obtained by observing beacons 1-hop away from the sensors. We show that under this simplified model the beacon deployment problem, i.e., to find the minimum number of beacons, is (still) NP-complete. We then provide an approximation algorithm with a logarithmic approximation ratio based on a set cover greedy heuristic; we study both offline and online versions of this approximation. Finally, we incorporate the localization accuracy into the beacon deployment strategy by explicitly specifying the Cramer Rao bound (CRB) as the condition. We verify the performance of the proposed algorithms using simulations, and show that accurate localization can be effectively achieved using the proposed algorithms.

The ad hoc network localization problem deals with estimating the geographical location of all nodes in an ad hoc network, focusing on those nodes that do not have a direct way (for example, GPS) to determine their own location. Proposed solutions to the ad hoc localization problem (AHLP) assume that nodes are capable of measuring received signal strength indication (RSSI) and/or are able to do coarse (sectoring) or fine signal angle-of-arrival (AoA) measurements. Existing algorithms exploit different aspects of such sensory data to provide either better localization accuracy or higher localization coverage. However, there is a need for a framework that could benefit from the interactions of nodes with mixed types of sensors. In this paper, we study the behavior of RSSI and AoA sensory data in the context of AHLP by using both geometric analysis and computer simulations. We show which type of sensor is better suited for which type of network scenario. We study how nodes using either, both, or none of these sensors could coexist in the same localization framework. We then provide a general particle-filtering framework, the first of its kind, that allows heterogeneity in the types of sensory data to solve the localization problem. We show that, when compared to localization scenarios where only one type of sensor is used, our framework provides significantly better localization results. Furthermore, our framework provides not only a location estimate for each nonanchor, but also an implicit confidence measure as to how accurate this estimate is. This confidence measure enables nodes to further improve on their location estimates using a local, iterative one-hop simple message exchange without having to rely on synchronized multiphase operations like in traditional multilateration methods.

In this paper, we present a mathematical model for a novel TCP congestion control approach called Receiver-Window Modification (RWM). RWM could be used with any of the present Active Queue Management (AQM) schemes, such as Random Early Detection (RED), Adaptive Random Early Detection (ARED), and BLUE queues to reduce congestion at the ingress and gateway routers. We extend the work done on stochastic modeling of RED-ECN gateways to RED-RWM gateways. Using this model, we provide a weak convergence theorem for the number of connections; thus leading the way to more relaxed configuration of RED-RWM gateway parameters. We use a Monte Carlo simulation method to compare outcomes of our mathematical model with those of NS2 simulation experiments.

Explicit congestion notification (ECN) and active queue management (AQM) Schemes such as random early detection (RED), adaptive random early detection (ARED) and BLUE queues have been proposed for TCP/IP networks to compensate network congestion. However, using ECN requires that ECN be supported by both TCP senders and receivers. This paper presents a novel AQM modification called receiver- window modification (RWM). RWM can be used together with RED, ARED and BLUE queues, to provide congestion avoidance in packet switched networks at ingress and gateway routers. RWM does not require modification to all end system TCP/IP stacks but can be solely implemented in routers. Our RWM scheme helps in reducing the average queue sizes of RED, ARED, BLUE and even ARED-ECN, BLUE-ECN and RED- ECN queues. By reducing the average queue sizes, RWM queues reduce the queuing delay resulting in significant improvements in one-way end-to-end packet delays and dropped packets. It is also shown that the performance of RED- ECN, ARED-ECN and BLUE-ECN queues is heavily dependent on the queues of the downstream routers. RWM modified queues in ingress or gateway routers are not influenced by the number and state of the downstream router as they will piggyback congestion information to the source in the next available acknowledgement packet. We carry out extensive ns2 simulations to show our results and to support our claims.

This paper describes a probabilistic approach to global localization within an in-door environment with minimum infrastructure requirements. Global localization is a flavor of localization in which the device is unaware of its initial position and has to determine the same from scratch. Localization is performed based on the received signal strength indication (RSSI) as the only sensor reading, which is provided by most off-the-shelf wireless network interface cards. Location and orientation estimates are computed using Bayesian filtering on a sample set derived using Monte-Carlo sampling. Research leading to the proposed method is outlined along with results and conclusions from simulations and real life experiments.

The paper describes research towards a system for locating users in a home environment requiring only a minimal wireless infrastructure. The only sensor reading used for the location estimation is the radiofrequency received signal strength indication (RSSI) measured by an RF interface (e.g., Wi-Fi). Location estimates are computed using Bayesian filtering on sample sets derived by Monte Carlo sampling. Wireless signal strength maps for the filter are obtained by a two-step parametric and measurement driven ray-tracing approach to account for absorption and reflection characteristics of various obstacles. Our trace driven simulations indicate that RSSI readings from a single access point in an indoor environment are sufficient to derive good location estimates of users.

The recent industry standard Bluetooth promises low cost replacement of communication cabling with moderate symbol-rate, short-range wireless links. The same specification also addresses the establishment of point-to-multipoint piconets and the interconnection of several of these piconets into scatternets, enabling Bluetooth to be used as a technology for realizing personal area networks. Establishing Bluetooth piconets requires nodes to discover each other by completing an inquiry phase. This paper investigates the inquiry phase; shows the shortcomings of the current inquiry procedure in multi node-PAN scenarios, and outlines and analyses a backwards Bluetooth compliant modification to accelerate Bluetooth inquiry. Extensive simulations comparing the original and the proposed Bluetooth inquiry schemes show improvements of more than an order of magnitude in device discovery times.

Wireless ad hoc networks are becoming increasingly important in today's world. The most challenging problem in conjunction with ad hoc networks is routing, i.e., the procedure in charge of determining the trajectory of packets traveling over the network. For large-scale ad hoc networks scalability of the routing approach is extremely important. One of the approaches to scale up ad hoc routing is geographical location based routing, which usually requires all nodes to be aware of their exact locations. In this paper, a new routing algorithm is proposed which requires only a subset of nodes to know their exact location forming location areas around these nodes. This paper outlines the LABAR (Location Area Based Ad Hoc Routing) routing protocol and provides simulation measurements on its average routing distance compared to the optimum shortest path.

Bluetooth is an open specification for short-range wireless communication and networking, mainly intended to be a cable replacement between portable and/or fixed electronic devices. The specification also defines techniques for interconnecting large number of nodes in scatternets, thus enabling the establishment of a mobile ad hoc network (MANET). While several solutions and commercial products have been introduced for one-hop Bluetooth communication, the problem of scatternet formation has not yet been dealt with. This problem concerns the assignment of the roles of master and slave to each node so that the resulting MANET is connected. We introduce two novel protocols for forming connected scatternets. In both cases, the resulting topology is termed a bluetree. In our bluetrees the number of roles each node can assume are limited to two or three (depending on the protocol), thus imposing low slave management overhead. The effectiveness of both protocols in forming MANETs is demonstrated through extensive simulations.

A systematic and automatic method for dynamically optimizing medium access control (MAC) protocol parameters is presented. Our meta-protocol approach is capable of performing on-line optimization of critical MAC parameters without knowing in advance what network conditions will arise or how they may fluctuate over time. Furthermore, this dynamic optimization is achieved without any centralized control or exchange of control messages between nodes. The power of the new technique is demonstrated by two examples. In a LAN environment, it outperforms traditional contention based MAC protocols that adjust retransmission probabilities, e.g., employing binary exponential backoff. In a synchronous multi-hop environment, it automatically converges to the proper transmission schedule assignments for the actual node density.

This paper presents ABROAD, an adaptive medium access control (MAC) protocol for reliable broadcast packet transmission in wireless networks. ABROAD incorporates a collision-avoidance handshake within each slot of a synchronous transmission schedule, allowing nodes to reclaim and/or reuse idle slots while maintaining bounded access delay. Thus, ABROAD provides worst-case performance guarantees while remaining adaptive to local changes in traffic load and node connectivity. We analyze the optimal worst-case performance of ABROAD, and show that there is a strict increase in the number of broadcast packets per second over a pure time division multiple access (TDMA) protocol. Extensive simulation confirms our analysis, and also demonstrates that ABROAD outperforms broadcast protocols based on reliable unicast packet delivery schemes, such as the IEEE 802.11 MAC standard.

This paper describes an integrated framework for selecting optimal call mixes (in a multimedia traffic scenario) by bandwidth degradation in a wireless cellular network, to maximize the revenue earned by the service provider. Each admitted call in our framework generates a revenue for the service provider based on the parameters of the call. The sum of the revenues generated by all admitted calls at a time is considered as the total revenue earned in a cell. By degradation, we mean that: (1) some channels can be taken away from ongoing calls that are assigned multiple channels and/or (2) newly admitted calls that require multiple channels get fewer than what they requested. To avoid removing more channels from calls than they could tolerate, we incorporate a new call attribute: the degradation tolerance, i.e, the number of channels a call can be degraded without sacrificing the acceptable level of quality. We also consider priorities over calls to influence the admission-degradation decision. Our analytical framework includes both static and dynamic scenarios. The dynamic case is enhanced with the ability to select the optimal call mix using incoming and departing handoffs, new calls, and call terminations in a recursive way. We also discuss how to accommodate non-real-time calls into our system. To evaluate the performance of the proposed scheme a discrete event simulation tool, that accommodates our dynamic framework built on a customized simulated annealing optimization function, has been developed. Simulation results demonstrate that not only does the proposed integrated admission-degradation framework maximize the total revenue earned in cells, but also handoff and new call blocking probabilities are reduced.

A systematic and automatic method to dynamically combine any set of existing MAC protocols into a single higher layer, or meta-MAC protocol, is presented. The new approach makes it possible to always achieve the performance of the best component protocol, without knowing in advance which protocol will match the potentially changing and unpredictable network conditions. Moreover, this dynamic optimization is entirely automatic and runs without any centralized control or any exchange of messages, using only local network feedback information. We describe the method and prove that the resulting meta-MAC protocol achieves optimal performance in a well-defined sense. Through simulation on different types of networks and with different component MAC protocols, we demonstrate that our simple and practical combination algorithm yields highly adaptive and scalable MAC solutions.