Indoor localization of mobile nodes is receiving great interest due to the recent advances in mobile devices and the increasing number of location-based services. Fingerprinting based on Wifi received signal strength (RSS) is widely used for indoor localization due to its simplicity and low hardware requirements. However, its positioning accuracy is significantly affected by random fluctuations of RSS values caused by fading and multi-path phenomena. This paper presents a convolutional neural network (CNN) based approach for indoor localization using RSS time-series from wireless local area network (WLAN) access points. Applying CNN on a time-series of RSS readings is expected to reduce the noise and randomness present in separate RSS values and hence improve the localization accuracy. The proposed model is implemented and evaluated on a multi-building and multi-floor dataset, UJIIndoorLoc dataset. The proposed approach provides 100% accuracy for building prediction, 100% accuracy for floor prediction and the mean error in coordinates estimation is 2.77 m.

Software-defined networking (SDN) abstracts and centralizes the network control functions in a software entity that runs on a server, known as SDN controller. The controller needs to respond to its controlled elements in a strictly timely manner, and the controller placement has a prominent effect on its response time. Originally, all SDN architectures assumed a physical wired connection between the SDN controller and its controlled elements, and the controller placement problem (CPP) has been only studied under such wired settings. Recently, novel SDN architectures have been proposed in which a direct wireless connection is assumed between the controller and its controlled elements. In this paper, we consider the 'wireless CPP,' when the link between the controller and the controlled element is wireless. Specifically, our contributions are as follows. First, we propose two joint controller placement and assignment formulations, assuming wired links between the controllers and their controlled elements; the first formulation considers an average response time constraint, whereas the second one considers a per-link response time constraint. Then, using chance- constrained stochastic programming (CCSP), we extend our formulation to the case when the links between the controllers and their controlled elements are wireless. Finally, we evaluate our joint placement and assignment schemes under various system parameters. Our results demonstrate the advantage of our joint scheme, in terms of reducing the required number of controllers, compared to a recent sequential assignment and placement scheme in the literature. They also show the ability of our CCSP-based scheme in probabilistically satisfying the controllers response time constraints.

One of the main characteristics in cognitive radios is situation awareness. By classifying the modulation schemes used in surrounding transmissions, a secondary user (SU) can identify the existing users in the system and adjust his/her transmission parameters accordingly. In this paper, we propose a multi-task learning (MTL) approach to recognize the modulation scheme used among a specific set of analog and digital modulations. This approach uses a deep convolutional neural network (CNN) to extract the necessary features in order to classify the different modulation schemes. The MTL is used to separately train the modulation classes that normally cause a considerable confusion and therefore improve the overall classification accuracy. Our results on the RadioML dataset show that the suggested architecture achieves higher overall classification accuracy compared to the recently proposed Convolutional, Long Short Term Memory (LSTM), Deep Neural Network (CLDNN). Our classification accuracy of 86.97% at 18 dB SNR outperforms the state-of-the-art with 5% relative improvement.

This paper describes an architecture for a node in a cognitive network that employs distributed learning and reasoning. We present the architecture and describe a method of distributed reasoning using an island genetic algorithm. We then formulate a channel allocation problem that is unique to the use of cognitive radio networks for dynamic spectrum access. We provide simulation results for the island genetic algorithm as applied to our channel allocation problem.