Direction of arrival (DoA) estimation is a common sensing problem in radar, sonar, audio, and wireless communication systems. It has gained renewed importance with the advent of the integrated sensing and communication paradigm. To fully exploit the potential of such sensing systems, it is crucial to take into account potential hardware impairments that can negatively impact the obtained performance. This study introduces a joint DoA estimation and hardware impairment learning scheme following a model-based approach. Specifically, a differentiable version of the multiple signal classification (MUSIC) algorithm is derived, allowing efficient learning of the considered impairments. The proposed approach supports both supervised and unsupervised learning strategies, showcasing its practical potential. Simulation results indicate that the proposed method successfully learns significant inaccuracies in both antenna locations and complex gains. Additionally, the proposed method outperforms the classical MUSIC algorithm in the DoA estimation task.

In this paper, we focus on the problem of the phase noise management when working at high frequencies close to Sub-Thz or THz bands. We consider the exponential approximation of the phase noise autocorrelation to insert new pilots for the discovery of the said autocorrelation that is helpful for receiver processings such as the Wiener filtering. We consider full-pilots symbols and sparse-pilots symbols in a DFT-s-OFDM chain for which we describe the allocation parameters to ensure a fair discovery taking into account the impact on the spectral efficiency and on the discovery latency. We propose a simple procedure between any transceiver and any receiver to agree on the good pilot scheme parameters.

The orthogonal frequency division multiplexing (OFDM) waveform presents some limitations in time-varying channels, experienced e.g., in high-mobility environment. In this paper, we introduce a new channel estimation approach to enhance OFDM system performance in high-Doppler channels. This novel approach consists in inserting a pilot in the delay-Doppler domain, without puncturing the data in the frequency-time domain. Therefore, it avoids pilot overhead when the channel is very selective. Simulation results show that the proposed method outperforms the standard techniques by achieving a higher data throughput in high-mobility scenarios.

We propose a new algorithm to identify a Wiener-Hammerstein system. This model represents a communication channel where two linear filters are separated by a non-linear function modelling an amplifier. The algorithm enables to recover each parameter of the model, namely the two linear filters and the non-linear function. This is to be opposed with estimation algorithms which identify the equivalent Volterra system. The algorithm is composed of three main steps and uses three distinct pilot sequences. The estimation of the parameters is done in the time domain via several instances of the least-square algorithm. However, arguments based on the spectral representation of the signals and filters are used to design the pilot sequences. We also provide an analysis of the proposed algorithm. We estimate, via the theory and simulations, the minimum required size of the pilot sequences to achieve a target mean squared error between the output of the true channel and the output of the estimated model. We obtain that the new method requires reduced-size pilot sequences: The sum of the length of the pilot sequences is approximately the one needed to estimate the convolutional product of the two linear filters with a back-off. A comparison with the Volterra approach is also provided.

Modern communication systems rely on accurate channel estimation to achieve efficient and reliable transmission of information. As the communication channel response is highly related to the user's location, one can use a neural network to map the user's spatial coordinates to the channel coefficients. However, these latter are rapidly varying as a function of the location, on the order of the wavelength. Classical neural architectures being biased towards learning low frequency functions (spectral bias), such mapping is therefore notably difficult to learn. In order to overcome this limitation, this paper presents a frugal, model-based network that separates the low frequency from the high frequency components of the target mapping function. This yields an hypernetwork architecture where the neural network only learns low frequency sparse coefficients in a dictionary of high frequency components. Simulation results show that the proposed neural network outperforms standard approaches on realistic synthetic data.

In this paper, we focus on the selection of a pilot pattern to track the phase noise in high frequency bands in a DFT-s-OFDM chain. By considering the Wiener filter at the receiver side to perform the tracking, we use the inner cost function of the said filter as the cost function for the pilot selection. To obtain this cost function, the phase noise autocorrelation is required. Therefore, we introduce a new mathematical approximation of the autocorrelation function of the practical 3GPP phase noise model. At first, this leads to an analytical expression of the Wiener filter coefficients. Then, the said coefficients allow us to obtain an analytical expression of the cost function. Thus, by means of this result, we are able to provide a pilot pattern that jointly satisfies a constraint on the pilot overhead and a constraint on the minimum performance of the Wiener filter.

We consider the following positioning problem where several base stations (BS) try to locate a user equipment (UE): The UE sends a positioning signal to several BS. Each BS performs Angle of Arrival (AoA) measurements on the received signal. These AoA measurements as well as a 3D model of the environment are then used to locate the UE. We propose a method to exploit not only the geometrical characteristics of the environment by a ray-tracing simulation, but also the statistical characteristics of the measurements to enhance the positioning accuracy.

On the one hand, sign-bit shaping is a popular shaping scheme where the conditional probability of the sign bit is made non-equiprobable. On the other hand, probabilistic amplitude shaping (PAS) is a popular coding scheme, to combine shaping and a bit-interleaved coded modulation (BICM), where the sign bit should not be involved in the shaping. Indeed, with the PAS scheme the sign bit is the parity bit, i.e., the output of the systematic error-correcting code. As a result, sign-bit shaping has been used with multilevel coded modulations rather than BICM. In this paper, we show that with minor modifications it is possible to use sign-bit-like shaping with a BICM. Simulation results are provided with the 5G NR LDPC BICM scheme.

Standard Markov decision process (MDP) and reinforcement learning algorithms optimize the policy with respect to the expected gain. We propose an algorithm which enables to optimize an alternative objective: the probability that the gain is greater than a given value. The algorithm can be seen as an extension of the value iteration algorithm. We also show how the proposed algorithm could be generalized to use neural networks, similarly to the deep Q learning extension of Q learning.

In the context of railway systems, the application performance can be very critical and the radio conditions not advantageous. Hence, the communication problem parameters include both a survival time stemming from the application layer and a channel error probability stemming from the PHY layer. This paper proposes to consider the framework of Markov Decision Process (MDP) to design a strategy for scheduling radio resources based on both application and PHY layer parameters. The MDP approach enables to obtain the optimal strategy via the value iteration algorithm. The performance of this algorithm can thus serve as a benchmark to assess lower complexity schedulers. We show numerical evaluations where we compare the value iteration algorithm with other schedulers, including one based on deep Q learning.

We consider a multi-agent system where agents compete for the access to the radio resource. By combining some application-level parameters, such as the resilience, with a knowledge of the radio environment, we propose a new way of modeling the scheduling problem as an optimization problem. We design accordingly a low-complexity solver. The performance are compared with state-of-the-art schedulers via simulations. The numerical results show that this application-oriented scheduler performs better than standard schedulers. As a result, it offers more space for the selection of the application-level parameters to reach any arbitrary performance.

This paper investigates the decoding of a remarkable set of lattices: We treat in a unified framework the Leech lattice in dimension 24, the Nebe lattice in dimension 72, and the Barnes-Wall lattices. A new interesting lattice, named$L_{3\cdot 24}$, is constructed as a simple application of the single parity check on the Leech lattice. The common aspect of these lattices is that they can be obtained via a single parity check or via the$k$-ing construction. We exploit these constructions to introduce a new efficient paradigm for decoding. This leads to efficient list decoders and quasi-optimal decoders on the Gaussian channel. Both theoretical and practical performance (point error probability and complexity) of the new decoders are provided.

We characterize the complexity of the lattice decoding problem from a neural network perspective. The notion of Voronoi-reduced basis is introduced to restrict the space of solutions to a binary set. On the one hand, this problem is shown to be equivalent to computing a continuous piecewise linear (CPWL) function restricted to the fundamental parallelotope. On the other hand, it is known that any function computed by a ReLU feed-forward neural network is CPWL. As a result, we count the number of affine pieces in the CPWL decoding function to characterize the complexity of the decoding problem. It is exponential in the space dimension$n$, which induces shallow neural networks of exponential size. For structured lattices we show that folding, a technique equivalent to using a deep neural network, enables to reduce this complexity from exponential in$n$to polynomial in$n$. Regarding unstructured MIMO lattices, in contrary to dense lattices many pieces in the CPWL decoding function can be neglected for quasi-optimal decoding on the Gaussian channel. This makes the decoding problem easier and it explains why shallow neural networks of reasonable size are more efficient with this category of lattices (in low to moderate dimensions).

We present new efficient recursive decoders for the Barnes-Wall lattices based on their squaring construction. The analysis of the new decoders reveals a quasi-quadratic complexity in the lattice dimension. The error rate is shown to be close to the universal lower bound in dimensions 64 and 128.

Point lattices and their decoding via neural networks are considered in this paper. Lattice decoding in ℝn, known as the closest vector problem (CVP), becomes a classification problem in the fundamental parallelotope with a piecewise linear function defining the boundary. Theoretical results are obtained by studying root lattices. We show how the number of pieces in the boundary function reduces dramatically with folding, from exponential to linear. This translates into a two-layer ReLU neural network requiring a number of neurons growing exponentially in n to solve the CVP, whereas this complexity becomes polynomial in n for a deep ReLU neural network.