Abstract

In buildings, a large chunk of energy is spent on heating, ventilation and air conditioning systems. One way to optimize their usage is to make them demand-driven depending on human occupancy. This paper focuses on accurately estimating the number of occupants in a room by leveraging multiple heterogeneous sensor nodes and machine learning models. For this purpose, low-cost and non-intrusive sensors such as CO 2 , temperature, illumination, sound and motion were used. The sensor nodes were deployed in a room in a star configuration and measurements were recorded for a period of four days. A regression based method is proposed for calculating the slope of CO 2 , a new feature derived from real-time CO 2 values. Supervised learning algorithms such as linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), support vector machine (SVM) and random forest (RF) were used on several different combinations of feature sets. Moreover, multiple performance metrics such as accuracy, F1 score and confusion matrix were used to evaluate the performance of our models. Experimental results demonstrate a maximum accuracy of 98.4% and a high F1 score of 0.953 for estimating the number of occupants in the room. Principal component analysis (PCA) was also applied to evaluate the performance of a dataset with reduced dimensionality.

Abstract

This paper focuses on the customized-wireless sensor node implementation of the classical least mean square (LMS) algorithm for the reduction in data-transmissions from the sensor nodes to the sink in internet of things (IoT) networks. This reduction, in turn, increases the battery life of the sensor node. The system was deployed in outdoor and indoor environments to read the ambient temperature and then perform the prediction of the sensed data in order to minimize the number of data-transmissions to the sink node. The utility of the proposed concept has been demonstrated using the measured data and the battery life is increased 2.64 and 2.53 times in indoor and outdoor environments, respectively.

Abstract

This paper proposes a distributed implementation of spatial modulation (SM) using cognitive radios. In distributed spatial modulation (DSM), multiple relays form a virtual antenna array and assist a source to transmit its information to a destination. The source broadcasts its signal, which is independently demodulated by all the relays. Each of the relays then divides the received data in two parts: the first part is used to decide which one of the relays will be active, and the other part decides what data it will transmit to the destination. An analytical expression for symbol error probability is derived for DSM in independent and identically distributed (i.i.d.) Rayleigh fading channels. The analytical results are later compared with Monte Carlo simulations. Further, an instantaneous symbol error rate (SER) based selection combining is proposed to incorporate the direct link between the source and destination with existing DSM. Next, DSM implementation is extended to a cognitive network scenario where the source, relays, and destination are all cognitive radios. A dynamic frequency allocation scheme is proposed to improve the performance of DSM in this scenario. The frequency allocation is modeled through a bipartite graph with end-to-end SER as a weight function. The optimal frequency allocation problem is formulated as minimum weight perfect matching problem and is solved using the Hungarian method. Finally, numerical results are provided to illustrate the efficacy of the proposed scheme.

Abstract

Cognitive radio offers a novel solution to over-come the problem of spectrum underutilization by providing spectral access to secondary users. The television (TV) bands are of particular interest for secondary usage due to their high penetration power and greater coverage. These licensed bands are occupied only in few regions while in most of the other regions they are unoccupied and are termed asTV white space(TVWS). The estimation of TVWS has mostly been done by usingthe statistical and empirical propagation models. In this paper,terrain data is incorporated into the estimation of TVWS and shown to improve the quantitative estimation of TVWS. Using the relevant transmitter information and terrain data in the Indian state of Telangana, the efficacy of the proposed approach is demonstrated. The performance of the proposed approach is compared to that of widely used H at a propagation model. It is shown that the accuracy in TV coverage estimation increases onan average by 45% while incorporating terrain data as compared to using only empirical propagation model. As area outside theTV coverage is TVWS, the increased accuracy in the estimation of TV coverage directly translates to improved accuracy in the estimation of TVWS, which in turn translates into more efficient use of spectrum and better interference management

Abstract

Noise power uncertainty is a major issue in detectors for spectrum sensing. Any uncertainty in the noise power leads to significant reduction in the detection performance of the energy detector and also results in a performance limitation in the form of SNR walls. In this paper, we propose an evidence theory (also called Dempster-Shafer theory (DST)) based cooperative energy detection (CED) for spectrum sensing. The noise variance is modeled as a random variable with a known distribution. The analyzed system model is similar to a distributed parallel detection network where each secondary user (SU) evaluates the energy from its received signal samples and sends it to a fusion center (FC), which makes the final decision. However, in the proposed DST-based CED scheme, the SUs sends computed belief-values instead of actual energy value to the FC. Any uncertainty in the noise variance is accounted for by discounting the belief values based on the amount of uncertainty associated with each SU. Finally, the discounted belief values are combined using Dempster rule to reach at a global decision. Simulation results indicate that the proposed DST scheme significantly improves the detection probability at low average signal-to-noise ratio (ASNR) in comparison to the traditional sum fusion rule in the presence of noise uncertainty.

Abstract

This paper focuses on evaluating the gains obtained through cooperative spectrum sensing in the real world while using cyclostationary-based mobile sensors. In cooperative sensing(CS), different secondary users (SUs) in a geographical neighbor-hood cooperate to detect the presence of a primary user (PU).Compared with single-user sensing, cooperation provides diversity gains in the face of multi path fading and shadowing. The effectiveness of CS is demonstrated by analyzing data acquired in two extensive field measurement campaigns. The first measurement campaign (MC-I) focuses on measurements at fixed locations,whereas the second measurement campaign (MC-II) focuses on a scenario where measurements are taken inside a moving car.These measurements are carried out for DVB-T channels in the Capital Region of Finland, which consists of urban and suburban environments. Hard decision rules such as OR,AND,and MAJOR-ITYand a soft decision rule such as sum of cyclostationary test statistics (SUM) are employed, and their detection performances are compared with a cyclostationary-based single-user detector. A performance parameter of relative increase in probability of detection (RIPD) is used to efficiently demonstrate the cooperation gain obtained relative to local sensing. It is shown that cooperation can significantly improve the performance of a sensor severely affected by fading and shadowing effects. Furthermore, it is shown that increasing the number of collaborating users beyond few users(five to eight) does not, in practice, bring significant improvement in terms of the expected RIPD. The performances of CS schemes evaluated from MC-I are also compared with the corresponding simulated CS results using empirical channel models and terrain data for the same experimental parameters. It is shown that the use of empirical or theoretical models may result in detection errors in practical conditions, and measurements should be used to improve the accuracy in such scenarios.

Abstract

Cepstral analysis has been widely used in audio and speech processing applications because of its ability to reveal periodicities in a signal. The presence of cyclic prefix (CP) in orthogonal frequency division multiplexing (OFDM) signals induces periodicities. Motivated by this, the paper focuses on cepstral analysis of OFDM signal. The distributions of cepstral coefficients are de-rived for two scenarios of noise only and OFDM signal in noise. Itis shown that the OFDM cepstrum is significantly different from the additive white Gaussian noise (AWGN) cepstrum and can be used to detect OFDM waveforms. It is also shown that the cepstrum of OFDM is rich in features and can be used to estimate OFDM parameters such as number of sub carriers and length of the CPin an OFDM symbol. These OFDM waveform parameters can be used to automatically recognize or classify different OFDM waveforms, which are important for cognitive radios, coexistence of heterogeneous networks and signal intelligence. Two schemes are proposed to detect OFDM based primary user (PU) signalsin cognitive radio systems. The distributions of the test statistics under the two hypotheses are established.Neyman–Pearson detec-tions trategy is employed.Algorithms for estimating the number of subcarrier and the length of the CP are proposed and their performances studied through simulations. Later the proposed schemes are extended to cooperative sensing scenario with multiple secondary users (SUs) and it is shown that the collaboration between them significantly improve the performance of the proposed cepstrum based detection and estimation schemes.

Abstract

Cooperative spectrum sensing (CSS) is one of the efficient scheme that helps in improving the spectrum sensing performance of a cognitive radio network. In this paper we propose a Dempster-Shafer theory (DST) based CSS for cognitive radio network. The fusion rule is the D-S combination rule which is well known for its ability to handle uncertainty and has been used in quite a different number of fields. Noise uncertainty, which is unavoidable in practical field, can gravely limit the detection performance of CSS. In this paper we specially investigate sensor nodes undergoing uncertainty in noise power when the detection scheme is based on energy of the received signal and use DS theory to minimize such effects. Simulation results reveal significant improvement in CSS gain as compared to previous traditional methods.