Rafael Kunst

Improving reinforcement learning algorithms for dynamic spectrum allocation in cognitive sensor networks

Cognitive Radio Networks enable a higher number of users to access the spectrum of frequency simultaneously. This access is possible due to the implementation of dynamic spectrum allocation algorithms. In this context, one of the main algorithms found in the literature is the reinforcement learning based approach called Q-Learning. Although been widely applied, this algorithm does not take into account accurate information about the behavior of users neither the channel propagation conditions. In this sense, we propose three improvements to the dynamic spectrum allocation algorithms based on reinforcement learning for cognitive sensor networks. Simulation results show that all the proposed algorithms allow allocating channels with up to 6dB better quality and 4% higher efficiency than Q-Learning.

On the impact of hybrid errors on mobile WiMAX networks

Burst and AWGN errors affect mobile WiMAX networks because of propagation conditions generally resulting from the mobility characteristics. Mobile WiMAX networks are inherently vulnerable to transmission errors due to propagation conditions such as multipath fading, shadowing, and Doppler spectrum. To ensure reliable communications, even in adverse physical conditions, errors must be detected and corrected by the receiver device. The traditional approach to deal with this problem is the employment of forward error correction techniques along with bit interleaving during the phase called channel encoding. In this article we propose an hybrid errors model where both burst and AWGN errors are considered. Moreover, we present an error sequence generator, used to simulate and evaluate the use of forward error correction techniques and bit interleaving applied to nomadic (fixed) and mobile WiMAX systems affected by hybrid errors. Simulation results show that using a hybrid error model better reflects the realistic behavior of RF channels, than models which consider only AWGN.

A self-adapting connection admission control solution for mobile WiMAX: Enabling dynamic switching of admission control algorithms based on predominant network usage profiles

WiMAX is a connection-oriented wireless network that provides QoS in metropolitan broadband communications. One important component in WiMAX QoS provisioning and management is the Connection Admission Control (CAC), which must be aware of the network conditions (e.g., user traffic demands and physical aspects). In our research, we define the association between a particular user traffic demand and a specific physical condition as a network usage profile. State-of-the-art proposals focus on optimizing CAC algorithms considering a single network usage profile; the adaptation of CAC algorithms when the predominant network usage profile changes is partially or fully neglected. In this article, we introduce a self-adapting CAC solution that, using a library of CAC algorithms, is able to switch the running algorithm according to the current network usage profile. The evaluation results, obtained through simulations, demonstrate that our self-adapting CAC solution is able to detect the changes on the predominant network usage profile. In addition, the results show how much different profiles can impact on the efficiency of CAC algorithms, thus confirming the need of switching the running CAC algorithm so that QoS can be guaranteed for the ongoing connections.

Improving QoS in multi-operator cellular networks

Providing high quality network access is challenging for network operators in the current static model used for allocating the spectrum of frequencies. Dealing with this challenge demands optimized resources allocation. One of the ways of providing this optimization is by allowing resources sharing among network operators which share the same geographical area. To allow and control the sharing of resources in such network scenarios, in this paper, a multilevel broker is presented to allow network operators to share their underutilized resources. This broker dynamically establishes a service level agreement that takes into account the quality of service requirements of the resources renters. A performance evaluation conducted in a scenario composed of multiple LTE-Advanced network operators shows that the implementation of the proposed architectures leads to more efficient allocation of underutilized network resources compared to two algorithms found in the literature.

A Resources Sharing Architecture for Heterogeneous Wireless Cellular Networks

The current static model used for allocating the spectrum of frequencies and the increasing demand for network resources imposed by modern applications and services may lead to a resources scarcity problem. Dealing with this problem demands optimized resources allocation. An alternative to provide this optimization is by allowing resources sharing among network operators. In this paper, an architecture is presented to encourage network operators to share their underutilized resources. A multilevel broker is proposed to control the resources sharing. This broker dynamically establishes a service level agreement that takes into account the quality of service requirements of resources renters and the cost of the resources. A performance evaluation shows that the implementation of the proposed architectures leads to better allocation of underutilized network resources.

Energy efficiency evaluation of the pulse shapes and modulation techniques for IR-UWB in WBANs

This paper evaluates the impacts of pulse shapes and modulation format on energy and power consumption for wireless body area networks (WBANs) based on ultra wideband (UWB). Microelectronic devices (e.g., sensors, transceivers, etc.) are components of this network, used to monitor biological signal activities in vicinity, over the body or by implants. An impulse radio (IR) is addressed for this purpose using UWB frequencies. The basic waveforms Gaussian pulse derivatives and Prolate Spheroidal wave function (PSWF) are presented to study the tradeoff between time- and frequency-limited signals as a possible solution to increase the energy efficiency. The total energy consumption is measured via simulations, considering the signal characteristics, the WBAN payload, the pulse position modulation (PPM), and on-off keying (OOK) schemes.

Proposal of an error sequence generator applied to the performance analysis of IEEE 802.16

RF channels are affected by time and frequency variant conditions caused by physical impairments that lead to the occurrence of errors during the transmission. These errors are basically of two types: burst errors and random errors. Simulating the behavior of RF channels affected by physical impairments has been subject of several investigations in the past years. Nevertheless, researches generally do not consider the occurrence of both errors at the same time. The current approach to simulate RF channel behavior may lead to imprecisions on the results. In order to deal with this limitation, this paper proposes an hybrid error sequence generator, which is able of generating both burst and random error sequences. The proposed error sequence generator is applied to a case study that aims to evaluate the performance of the channel encoder of the IEEE 802.16 networks. The outcomes of the experiments show that using an hybrid model reflects better the realistic behavior of RF channels.

Analysis of WiMAX bandwidth allocation mechanism considering physical conditions

The increasing development of mobile applications has changed the traditional demands of computer networks, which must now provide ubiquitous broadband wireless communication. In this context, IEEE 802.16 standardizes the physical and medium access control layers of the technology, aiming to allow network access in both nomadic and mobile scenarios. In this work we analyze the impact caused by overhead on the bandwidth allocation mechanism of WiMAX networks. Our main focus is to evaluate the number of SSs that can be served, and the influence of adaptive modulation and coding configuration. Finally, we discuss the importance of considering physical impairments in the process of designing a broadband wireless access network.

Improving network resources allocation in smart cities video surveillance

Abstract Smart cities is an emerging concept that allows the usage of technology to improve the quality and efficiency of services delivered to the population of a given city, considering a vast application domain. Some applications, like smart surveillance using Internet of Things sensors demand the transmission of high amounts of traffic with strict quality of service, what may not be available at all times. One solution to deal with this traffic overload is to implement the concept of resources sharing. Considering this context, in this article, a heterogeneous network scenario is considered to implement a multi-purpose real time video surveillance application which can be applied to both smart cities and military surveillance such as borderline security. The proposed solution considers three crucial aspects: the QoS requirements of the real time video application; the cost-benefit of the spectrum allocation; and the time constraints involved in a vertical handover operation to avoid interference. The main contribution of the article are: (I) inclusion of a multilevel resources broker in the Smart Cities architecture, (II) a QoS-enabled solution for video surveillance to provide borderline security, and (III) a low overhead control mechanism to avoid interfering with the legitimate network traffic. Results obtained via simulations show that after the proposed solution is implemented delay and jitter are kept below their specific thresholds and therefore the QoS is guaranteed considering the coexistence of up to 500 IoT sensors. Nearly 70% of the traffic is accommodated either in IEEE 802.11 or IEEE 802.22 networks, which offer relatively low cost resources. These results are also compared with two state of the art related works and show that the proposed solution performs better than the related work in the analyzed scenarios. Finally, with respect to the time constraints, the implementation of the proposed solution leads to fast vertical handover, attaining times up to 46% lower than the maximum allowed handover duration.

A cognitive algorithm for traffic steering in LTE-LSA/Wi-Fi resource sharing scenarios

The wireless network traffic is expected to overload the existing licensed spectrum by 2020. One method to deal with this traffic overload is to access unlicensed and shared spectrum bands using an opportunistic approach. Licensed Shared Access (LSA) allows incumbent users to provide temporary access to its spectrum resources. However, licensees must perform traffic steering to vacate the band without causing interference, whenever the incumbent requires. In this paper, a cognitive algorithm is proposed to take in advance decisions to promptly create a list of traffic steering routes whenever an unscheduled evacuation is demanded. This solution aims at guaranteeing the QoS and seamless connectivity during traffic steering. A performance evaluation conducted in a scenario composed of one LTE-LSA and three Wi-Fi network operators demonstrates that the proposed solution fulfills the time required by the unscheduled evacuation as well as guarantees the QoS and seamless connectivity of evacuees.