The fast expansion of Industrial Internet of Things (IIoT) applications demands energy-efficient and high-speed communication technologies. 5G, which relies on Orthogonal Frequency Division Multiplex (OFDM), is a promising solution, but its energy consumption remains a challenge. In OFDM, pilot symbols are used for channel estimation, and a cyclic prefix (CP) is added to mitigate inter-symbol interference, both of which increase energy consumption and reduce transmission efficiency. To address this issue, we propose a CP-Free OFDM transmission scheme combined with a new blind channel estimation method adapted for every M-QAM modulation. Our approach employs an iterative algorithm to eliminate the need for a CP and pilot symbols, reducing energy consumption and improving spectral efficiency.

The use of a reliable and accurate localization system is essential to ensure the smooth integration of visually impaired people into society and making their daily lives easier. Many of the proposed systems are based on the use of Machine Learning (ML) where increasingly complex models are introduced. However, this is not always necessary given that comparable performance can be achieved with simpler models by better understanding and adjusting key model parameters that impact model performance. In this paper, we present an indoor localization framework that applies a simple 1D Convolutional Neural Network (CNN) on a real collected LoRa dataset. We concentrate our work on studying and analyzing the impact of essential parameters and data criterion on the efficiency and performance of the system. We mainly study the impact of the LoRa frequency, the impact of the distribution of training points and the impact of the quality of collected fingerprints. Several conclusions have been drawn showing the correlation between localization performance and the specific features of propagation and model training.

Data imbalance is a major challenge in vehicle testing and autonomous driving datasets, where rare but critical driving scenarios (such as emergency braking or extreme weather conditions) are often underrepresented. This imbalance can introduce bias in machine learning models and compromise the accuracy of vehicle behavior predictions. In this paper, we propose a hybrid approach that combines Conditional Wasserstein Generative Adversarial Networks (CWGAN-GP) for oversampling imbalanced data and ensemble learning techniques (Random Forest (RF) and Extreme Gradient Boosting (XGBoost)) for vehicle behavior prediction. The GAN component generates realistic driving data to rebalance under-represented conditions, while the RF and XGBoost models leverage both real and synthetic data to enhance prediction accuracy. Experimental results in real-world vehicle test datasets show that our approach significantly improves model performance compared to conventional methods, achieving a 80% increase in classification accuracy on the minority class and a 30% improvement in AUC (Area Under the ROC Curve), a metric used to evaluate binary classification models. Furthermore, our method was validated on the open source SECOM dataset, which is known for its extreme imbalance. This hybrid methodology offers a promising solution to improve machine learning models in autonomous vehicle testing and intelligent transportation systems.

In recent years, several of the latest communication technologies enable Round Trip Time (RTT) for positioning in addition to Received Signal Strength (RSS). It is expected to improve consistency of the observable locations. This approach has been gaining support from several companies such as Google, which introduced this feature in the Android system. As a result, RTT estimation is now available in several recent off-the-shelf devices, opening a wide range of new approaches for estimating location. RTT is most of the time associated with trilateration solutions. However, few works exist assessing the feasibility and accuracy of RTT as an input for fingerprinting based positioning. In this paper, we try to contribute to fill this gap by investigating the performance of ranging fingerprint indoor LoRa based localization solution in terms of accuracy and positioning errors compared to RSS. The ranging is estimated from RTT measurements. Results reveal RTT or ranging is an excellent input for fingerprinting solutions and its fusion with RSS achieve a very good accuracy in both Line-of-Sight (LOS) and Non-Line-of-Sight (NLOS) scenarios.

To localize objects in an indoor environment, several methods are used such as trilateration and fingerprinting. These methods are based on the Received Signal Strength Indicator (RSSI), which is very sensitive to propagation factors indoor due to the existence of different static and dynamic obstacles. The use of RSSI causes the problem of missing data because the RSSI transmitted by an access point is not correctly received by the object sensors. To deal with the problem of missing data, researchers propose several completion methods, nevertheless no definitive solution has been brought. Therefore, we propose in this paper to use Generative Adversarial Networks for data recovery in order to be used efficiently for objects localization. The results based on the simulation data show an improvement of the localization accuracy compared to the classical methods. The simulations were performed with 10% and 50% of missing data and improved of 29.19% and 37.51 % respectively the localization accuracy.

This paper focuses on the problem of indoor localization based on radio signal measurements. The results presented are based on real measurements performed in Beirut and Paris. These measurements were performed in visibility and non-visibility contexts, using the Receiver Signal Strength Indicator (RSSI) of LoRa signals, by precisely controlling the number of gateways and the number of measurement points. A localization algorithm, based on a Multi-Layer Perceptron (MLP) neural network, was proposed and it was shown that it was possible to localize the measurement points accurately. The experimental results have shown a probability of localization greater than or equal to 98% with an accuracy of less than 1m. These results validate the stability of the RSSI measurements and confirm the possibility of using them for indoor localization.

As Sparse code multiple access (SCMA) has proved to be a fascinating research in order to meet the requirements of future wireless communication systems. To reach high power efficiency, wireless communication systems are equipped with high power amplifiers (HPAs). In this paper, we investigate the effects of distortions due to high power amplifiers (HPA) nonlinearities. We study the performance of amplified SCMA systems, in terms of bit error rate (BER). Message passing algorithm (MPA) is considered for SCMA detectors. BER performance is derived and evaluated for Additive White Gaussian Noise (AWGN) and Rayleigh fading channels. Numerical results and comparisons are provided for several system parameters, such as the input back-off (IBO). Indeed, we propose a new distortion cancellation technique based on feed-forwarded neural networks (FNNs) to restore the system performance via eliminating the HPA nonlinearities at transmitter and receiver sides. It is confirmed that the proposed pre-distorter and post-distorter with neural network exhibit a good performance improvement of quality of the transmission. Specifically, post-distortion based on NNs shows a better BER performance, which is almost close to the one of the linear system.

With expeditious development of wireless communications, location fingerprinting (LF) has nurtured considerable indoor location-based services (ILBSs) in the field of the Internet of Things (IoT). For most pattern-matching-based LF solutions, previous works either appeal to the simply received signal strength (RSS), which suffers from dramatic performance degradation due to sophisticated environmental dynamics, or rely on the fine-grained physical layer channel state information (CSI), whose intricate structure leads to increased computational complexity. Meanwhile, the harsh indoor environment can also breed similar radio signatures among certain predefined reference points (RPs), which may be randomly distributed in the area of interest, thus mightily tampering the location mapping accuracy. To work out these dilemmas, during the offline site survey, we first adopt autoregressive (AR) modeling entropy of CSI amplitude as location fingerprint, which shares the structural simplicity of RSS while reserving the most location-specific statistical channel information. Moreover, an additional Angle-of-Arrival (AoA) fingerprint can be accurately retrieved from the CSI phase through an enhanced subspace-based algorithm, which serves to further eliminate the error-prone RP candidates. In the online phase, by exploiting both CSI amplitude and phase information, a novel bivariate kernel regression scheme is proposed to precisely infer the target's location. Results from extensive indoor experiments validate the superior localization performance of our proposed system over previous approaches.

Sparse code multiple access (SCMA) has attracted growing research interests in order to meet the targets of the next generation of wireless communication networks. Since it relies on non-orthogonal multiple access (NOMA) techniques, it is considered as a promising candidate for future systems that can improve the spectral efficiency and solve the problem of massive user connections. In this paper, the basic concept of SCMA is introduced, including SCMA encoding, codebook mapping, and SCMA decoding. The major challenge of SCMA is the very high detection complexity. Then, a novel strategy for blind decoding based on convolutional neural networks is proposed. Through simulations, we showed that our proposed scheme outperforms conventional schemes in terms of both BER and computational complexity, where 0.9 dB improvements can be achieved.

Sparse code multiple access (SCMA) and polar codes (PC) are two promising candidates for Future communication systems since they are capable of achieving high system capacity. In this paper, we develop a novel detection and decoding scheme for SCMA systems combined with channel coding candidate polar codes. First, we propose a separate detection and decoding (SDD) receiver for uplink communications. Then, we introduce a joint detection and decoding (JDD) receiver scheme. The investigation of system receiver is decomposed on message passing algorithm (MPA) based SCMA multiuser detection and soft cancellation (SCAN) algorithm based polar codes decoder. The separate and joint schemes are studied over additive white gaussian noise (AWGN) channels. JDD scheme yields a better performance gain. Moreover, the joint scheme has a lower computational complexity compared to the separate one. Numerical results show that when polar code length polarN= 1024 and R = 1/2, under system loading 150%, JDD outperforms the SDD 1.8dB at BER = 10-2and 3.3dB at BER = 10-6over AWGN channels.

This paper presents a localization algorithm based on the identification of a linear expression connecting a vector of mean received powers to a vector of Cartesian coordinates. The linear expression is based on a Taylor expansion of the received power collected on a set of measurement points. The terms of the Taylor expansion are then integrated in a transfer matrix able to predict a vector of received power from a composite vector of Cartesian coordinates. Using the pseudo inverse of this matrix it is then possible to find the Cartesian coordinates of any unknown reception point, from the vector of mean received powers at this point.

Location Fingerprinting (LF) is a promising localization technique that enables many commercial and industrial Location-based Services (LBS). In this paper, a Channel Impulse Response (CIR) based indoor localization system is proposed. To fully exploit the most location-specific multipath information, we first conduct a power-based time tap filtering for the received CIR measurements. Furthermore, we experimentally observe that the filtered CIR data exhibits a multivariate circularly-symmetric Gaussian feature, which hints that fingerprinting positioning can be implemented by using a more accurate probabilistic method with a less computational complexity. In the online position determination phase, the Kullback-Leibler Distance (KLD) is adopted to quantify the similarities between the received measurements of target and the fingerprint database. Afterwards, we employ a probability kernel based regression approach to accurately infer the estimated target's location. Through extensive experiments performed on CRAWDAD database, the efficiency of our proposed scheme is validated.

This paper presents a comparison study of different similarity metrics used for RSSI fingerprint based indoor localization. These metrics are used for nearest neighbor search which is a crucial step in fingerprint localization system. Including Euclidean distance, Manhattan distance and Gauss distance, the present study compares the localization error respect to a proposed parameter named “error density”. This latter is related to the location error and the size of the studied area. In addition, different methods of combining the locations of neighbors have been introduced to estimate the current position and their performances have been compared. Extensive implementation details are discussed and simulation is conducted to compare them. Obtained results show that the Kernel method combined with the weighted average exhibits the localization accuracy on the studied dataset.