Applications like multi user and distributed MIMO can benefit from a sparse transmission scheme due to a reduced inter-cell and inter-cluster interference. In this paper, the so-called start stop bit method, which presents a new type of transmission with sparse resource usage is studied. Several enhancements are analyzed such as i) transmitting multiple stop bit sequences, ii) adding sparse stop bit areas representing artificial fractional time frequency shifts, iii) adding a given MCS over the stop bit areas, and iv) compressing the permutation matrix needed for multiple stop bit sequences. Bit error rate curves of the proposed transmission scheme are analyzed and suitable parameter sets are provided that lead to the highest throughput per set of resource elements for a given sparsity level in comparison with a conventional OFDMA system.

The application of artificial intelligence and machine learning (AI/ML) into the wireless physical layer is under discussion at 3GPP. Channel state information (CSI) prediction is among the sub use cases being studied. In this work, we propose an AI/ML CSI predictor that aims to compensate the scheduling delays at the base station. The AI/ML CSI predictor operates at the user equipment side and generates the channel reporting based on its prediction. Our AI/ML CSI predictor is designed for the intended prediction time, e.g., 5 ms, by collecting a few past measurements at the input. Our architecture is flexible regarding the number of physical resource blocks and can be used by all user equipments within the cell. Our results show that the proposed AI/ML CSI predictor has the 90 % normalized squared error performance around −13 dB and less than 1.4 % of the predicted eigenvectors have a squared generalized cosine similarity below 0.9, which is much better than zero order hold.

Machine learning (ML) applications for wireless communications have gained momentum on the standardization discussions for 5G advanced and beyond. One of the biggest challenges for real world ML deployment is the need for labeled signals and big measurement campaigns. To overcome those problems, we propose the use of untrained neural networks (UNNs) for MIMO channel recreation/estimation and low over-head reporting. The UNNs learn the propagation environment by fitting a few channel measurements and we exploit their learned prior to provide higher channel estimation gains. Moreover, we present a UNN for simultaneous channel recreation for multiple users, or multiple user equipment (UE) positions, in which we have a trade-off between the estimated channel gain and the number of parameters. Our results show that transfer learning techniques are effective in accessing the learned prior on the environment structure as they provide higher channel gain for neighbouring users. Moreover, the proposed UNN channel state information (CSI) estimators are under-parameterized and can further enable low-overhead CSI reporting.

One of the essential challenges for MIMO or massive MIMO and cooperative multipoint systems is obtaining accurate channel state information (CSI) at the base station side, as the related closed loop MIMO precoders are sensitive to channel aging effects. For that reason, channel prediction is often seen as one of the main enablers to maintain a large part of the performance gains achievable with ideal CSI. It has been claimed that Kalman filtering provides an upper bound for channel prediction performance. In line with this, for real world radio channels alternative channel prediction methods based on parameter estimation achieved comparably worse prediction results. Here, we propose a so called profiling solution as a novel parameter estimation method, which promises to improve the prediction horizon by a factor of two to three compared to Kalman filtering based on autoregressive models. This is indicated by first evaluations based on a real world radio channel measurement in the NOKIA campus in Munich.

Massive multiple input multiple output (MIMO) transmission and coordinated multipoint transmission are candidate technologies for increasing data throughput in evolving 5G standards. Frequency division duplex (FDD) is likely to remain predominant in large parts of the spectrum below 6 GHz for future 5G systems. Therefore, it is important to estimate the downlink FDD channels from a very large number of antennas, while avoiding an excessive downlink reference signal overhead. We here propose and investigate a three part solution. First, massive MIMO downlinks use a fixed grid of beams. For each user, only a subset of beams will then be relevant, and require estimation. Second, sets of coded reference signal sequences, with cyclic patterns over time, are used. Third, each terminal estimates its most relevant channels. We here propose and compare a linear mean square estimation and a Kalman estimation. Both utilize frequency and antenna correlation, and the later also utilizes temporal correlation. In extensive simulations, this scheme provides channel estimates that lead to an insignificant beamforming performance degradation as compared to full channel knowledge. The cyclic pattern of coded reference signals is found to be important for reliable channel estimation, without having to adjust the reference signals to specific users.

The ongoing standardization within 3GPP for the so called new radio (NR) system has identified massive multiple-input multiple output (MIMO) transmission, also called full dimension MIMO, as one of the main contributors to higher spectral efficiency for the mobile broadband case. In particular for radio frequencies below 6 GHz, channel estimation has to be supported in frequency division duplex (FDD) as well as time division duplex (TDD) operation. In TDD we may obtain downlink channels by estimating uplink channels, assuming reciprocity. For FDD, codebook based design as well as some type of explicit feedback is under discussion. Separately, there are also ongoing discussions of the question if massive MIMO in combination with FDD is a reasonable choice at all. Here we highlight some of our recent results obtained within several 5G research projects. To our understanding they overcome some of the inherent limitations of massive MIMO for FDD. As indicated by simulations, the resulting concept enables a grid of beam (GoB) and reference signal design with a reasonable downlink reference signal overhead of around 10 percent, together with reasonable feedback overhead of several hundred kbit/s per UE. Such a design attains around 90 percent of the massive MIMO system performance with ideal channel state information.

A massive multiple-input, multiple-output (mMIMO) downlink is considered with fixed wideband beams known as Grid of Beams (GoBs). The radio channel of a typical urban macro scenario will be shown to be sparse, and in combination with a linear maximum ratio combining beamforming method is used at the user equipment. Simulation results for urban macro cells show that reasonable spectral efficiencies are achieved with a moderate number of relevant channel components (RCCs) and accordingly limited feedback overhead for reporting of channel state information (CSI).

Future mobile radio systems, like 5G, are setting extremely demanding targets with respect to the number of served users, data rates, and latency etc. Advanced techniques such as massive MIMO and joint cooperation over several distributed radio stations are being developed to achieve these targets. Channel prediction has been deemed to be a potential main enabler for these techniques. Here, we exploit the correlation present in the multi-dimensional massive MIMO channel and apply HOSVD-based techniques to improve the prediction of weak channels by devising a denoising step before the state-of-the-art channel predictor. We show that at low signal-to-noise ratios and prediction horizons of 1 ms, there is a gain of more than 7 dB in prediction performance.

Joint transmission coordinated multipoint (JT CoMP) has been identified as a potential differentiator for future 5G radio systems due to its superior interference mitigation capabilities. Further, the combination with massive multiple-input-multiple-output (mMIMO) when a set of fixed grid of beams is used at eNodeB results in sparse overall channel matrices with a relatively low number of relevant channel components, which reduces the feedback overhead for reporting of channel state information (CSI). Although JT CoMP faces several challenges from synchronization to CSI outdating, in this paper the focus will be on complexity reduction of the precoding over large clustered cells utilizing massive MIMO. It will be derived how the typical sparse number of relevant channel components per user equipment can be exploited to reduce the number of floating point operations (FLOPs) by a factor of ten compared to state-of-the-art solutions for the calculation of the Moore Penrose pseudo inverse of the channel matrix.

For future 5G systems significant performance benefits are expected from massive MIMO, especially in combination with tight inter-cell cooperation including joint-transmission using cooperative multi-point transmission. Most massive MIMO evaluations concentrate on the base station side with the goal to achieve high spectral efficiency by MU-MIMO or large coverage by strong beamforming gains. Especially for the, here interesting, below 6 GHz RF-bands user equipment (UE) sided analysis is typically limited to four or mostly eight antenna elements per UE, which can be justified by the limited space to place more antenna elements as well as the related UE complexity. At the same time, there would be many benefits from UE-sided beamforming, ranging from improved channel estimation and prediction accuracy, effective interference suppression up to coverage and spectral efficiency gains on the system level. In our previous work, we have already proposed virtual beamforming for channel estimation and prediction, while here we extend the concept to user data transmission over virtually generated beams. Virtual beamforming directly applied to user data can be very inefficient, a challenge we overcome by parallel transmission over a set of coded virtual beams.

Massive MIMO (mMIMO) antenna array and Grid of Beams (GoB) are one of the key components for 5G to achieve high spectral efficiency and coverage. System level simulations are necessary to optimally intergrate the components of 5G system. Simulations with mMIMO antenna array involve handling large matrices, which results in increased simulation time typically upto several days depending on the size of the antenna array. However, GoB are fixed wideband beams and the receiver sees only the effective channel which is a combination of the channel and the GoB. Therefore, for system level simulations it is sufficient to general only the effective channel. In this paper, we propose a simulation model to directly generate the effective channel based on the 3GPP 3D channel model. It is shown that for the simulation of urban macro cells scenario specified in TR 38.913, the computational effort can be reduced by a factor of 1000. Furthermore, simulations show that very high SINR and user throughput values are achievable through GoB.

In November 2015 the latest world radio conference took place and as a result the below 6 GHz RF bands have become more important for 5G in the foreseeable future. In the 5GPPP project Fantastic5G options for higher spectral efficiency based on massive MIMO, joint transmission cooperative multi point and interference mitigation are being investigated. Most of these concepts have been already developed in previous projects like Artist4G or METIS, but now it is important to integrate these into a flexible framework supporting a rudimentary phase I as well as a future proof phase II system. Interference mitigation might become the main differentiator to LTE and the inevitable enablers will be shortly highlighted. According to current results the novel 5G system has the potential to increase spectral efficiency by roughly a factor of ten compared to a 4x2 LTE system, thereby significantly outperforming also latest LTE results for full dimension MIMO.

Conditions are derived on the interference channel with single-antenna nodes to achieve maximal degrees-of-freedom with only two symbol extensions. The conditions for line-of-sight channels involve choosing only the spacing between two subcarriers of an orthogonal frequency division multiplexing (OFDM) scheme. For 3 user pairs an upper bound on the sum-rate of interference alignment is approached arbitrarily closely.

Future cellular mobile radio systems - sometimes called beyond 4G (B4G) - will have to provide by factors higher network performance. Promising concepts like interference mitigation based on joint cooperative multipoint transmission (JT CoMP), multiple input multiple output (MIMO) or - more recently - massive MIMO depend on accurate and reliable channel state information. With so called model based channel prediction (MBCP) we proposed a powerful but challenging means, in theory promising to outperform state of the art Wiener or Kalman filtering. Here we analyze some important aspects of MBCP and suggest suitable enhancements for channel and parameter estimation of multi path components. Based on ray tracing tools and measurements the high potential of enlarged measurement bandwidth or virtual beamforming are being assessed. More work will be needed to understand the full potential and limitations of MBCP, being currently conducted in the EU funded project METIS.

Cooperative multipoint transmission (CoMP) mitigates strong inter cell interference for increased capacity in future networks and is a study item for 3GPP LTE advanced. We presented already the so called partial CoMP framework (pCoMP) as an efficient clustering approach, allowing to serve almost 90% of all user equipment (UE) user centric. We extended this framework to exploit multiple antenna elements at each of the cooperating cells for a powerful interference floor shaping technique, relying on wideband beamformers with variable TX power as well as antenna tilting and a joint precoding on top of that beams. Ray tracing and system level simulations indicate that SINRs of more than 20 dB are possible for a high number of users and in case of full reporting all channel components. Here we analyze the best trade off between more realistic partial reporting and achievable performance, e.g. a signaling overhead reduction of about 30% does not influence the performance significantly.

Mobile communication systems have evolved over the past decades and each new generation brought new experience to the users enabled by technology innovations, while keeping some well established principles from previous generations. This trend continued up to LTE (Long Term Evolution) Advanced, the predominant 4thgeneration system which has just been standardized in 3GPP and is being rolled out soon. How will this trend continue to future systems which will be deployed in some 10 years from now which will be advanced enough to be called “Beyond 4G” (B4G)? This article presents how such B4G systems will look like and some key technologies they will rely on including versatile numerology, massive virtual MIMO from many base stations, both centralized and distributed architectures using fiber optics as backbone, advanced interference mitigation, cognitive self organization, and wideband RF radios.

Cooperative multi point transmission (CoMP) is the topic of different investigations regarding LTE-advanced networks to fulfill future capacity requirements by mitigating inter cell interference. If joint transmission is performed by several sectors to their assigned UEs, the interference within this cooperation area (CA) is suppressed completely, but still existent between different CAs. An user centric assignment counters that inter CA interference, but is hard to realize because of the low probability to find a group of UEs with an identical set of strongest sectors. The so called `partial CoMP' (pCoMP) framework circumvents this limitation and is presented in this article with focus on intra CA interference. Dependencies between different parameters of pCoMP and the resulting intra CA interference show a trade off between needed system resources and remaining interference. System level simulations regarding intra and inter CA interference yield gains similar to scenarios using user centric assignment, but with a penetration rate of more than 90% of all users.

In the context of long term evolution (LTE) Advanced networks so called coordinated multi point transmission (CoMP) - where transmission is done jointly from several cells to user equipment (UE) - has gained a lot of interest lately. To benefit from CoMP by interference mitigation, UEs have to be served user centric to really leverage potential performance gains like increased capacity or spectral efficiency. This leads to a severe challenge how to assign user groups (UG) to cooperation areas (CA) in a user centric way because UGs must be found with an unique set of strongest cells. Here the so called partial CoMP concept is proposed and analyzed, which allows to serve up to 90% of all UEs in a user centric manner by their 3 strongest cells. The main achievement is to reach the significant performance gains of user centric cooperation for a practical and easy to realize system.

This paper presents an adaptive codebook design method, which introduces a transition probability based codeword sorting scheme in an effort to reduce the feedback overhead by exploiting the temporal correlation of the channel. The adaptive codebook sorts the codewords in the increasing order of the chordal distance with respect to the initial codeword, which is the codeword chosen at the previous time step. The adaptive codebook method in conjunction with the lossless data compression scheme requires much less feedback overhead compared with currently available techniques, e.g., the recursive codebook scheme, at the cost of slightly increased memory requirements.

Multicellular radio systems are often limited due to the presence of cochannel interference. Proposed physical layer concepts, e.g. coordinated joint transmission and interference rejection combining, try to strengthen the signal while combating the interference. However, the performance may be limited by the available channel knowledge. We provide a concept for multi-cell channel estimation in the downlink applicable for for both physical layer concepts. This concept uses virtual pilots based on block-orthogonal sequences, e.g. Hadamard.

A specially favorable concept for interference reduction in future mobile radio systems consists in the application of joint detection in the uplink and joint transmission in the downlink. Concerning joint detection both optimum nonlinear and low complexity suboptimum linear signal processing algorithms are well known. This is not the case for joint transmission, where up to now most research work focuses on linear signal processing algorithms, which are dual to the ones applied for joint detection. In the present paper a novel signal processing algorithm for joint transmission, termed transmit nonlinear zero forcing, is presented. The main characteristic is the application of an unconventional multiply connected quantization scheme, which allows to improve the energy efficiency, i.e., for data symbols to be transmitted over channels with high attenuations without introducing excessive interferences. Especially in scenarios with large gain differences of the channels between the access point and individual mobile terminals said improved energy efficiencies allow a significant reduction of the required transmitted powers, a feature mitigating intercell interference in cellular systems and thereby enhancing system capacity. In practical systems such gain differences, due to shadowing and largely different distances between the access point and the mobile terminals, prevail.

Modern wireless communication standards like HiperLAN/2 or IEEE802.11a provide a high degree of flexibility on the physical layer (PHY) allowing the data link control (DLC) layer to choose transmission parameters with respect to the currently observed link quality. This possibility of so-called link adaptation (LA) is a key element for meeting quality of service (QoS) requirements and optimizing system performance. The decisions made by the LA algorithm are typically based on a prediction of the packet error rate (PER) implied by a certain transmit parameter setting. With most existing LA schemes, this prediction is based on the observed signal-to-noise ratio (SNR) or carrier-to-interference ratio (CIR), respectively. For frequency selective channels, however, the SNR alone does not adequately describe the channel quality. It is shown in this paper that LA schemes based on an accurate description of the momentary channel status can outperform conventional schemes with respect to throughput and reliability. Based on a novel "indicator" concept, schemes for obtaining an accurate prediction of the PER by evaluation of the channel transfer function are presented and evaluated by software simulation.

Adaptive modulation, i.e. subcarrier-specific selection of the modulation scheme, is a well-known technique for enhancing the performance of multi-carrier systems. The adaptation to the frequency selective characteristics of the channel leads to considerable SNR gains for transmission over non-line-of-sight multipath channels. The actual assignment of modulation schemes to the individual subcarriers must be known to both the transmitter and the receiver. Conventionally this is achieved by exchanging information about the assignment over the communication link. i.e. a certain amount of signalling overhead is introduced. In this paper a signalling-free approach for reconstructing the assignment of modulation schemes in the receiver is presented. The decision about the modulation scheme used on each individual subcarrier is incorporated into the process of channel decoding.

Modern wireless communication standards like HiperLAN/2 or IEEE 802.11a provide a high degree of flexibility on the physical layer (PHY) allowing the data link control (DLC) layer to choose transmission parameters with respect to the currently observed link quality. This possibility of so-called link adaptation (LA) is also a key element for meeting quality of service (QoS) requirements of higher layers. Most LA algorithms take an estimate of the signal-to-noise ratio (SNR) or carrier-to-interference ratio (CIR), respectively, as the only input from the PHY layer. For frequency selective channels, however, the SNR alone does not adequately describe the channel quality. It is shown in this paper that system performance can be significantly improved if the LA scheme is based on a more detailed description of the channel situation. Methods for selecting the PHY mode based on the estimated channel transfer function are presented and compared to conventional algorithms.

Cellular networks and wireless local area networks (WLAN) are being widely deployed for mobile radio communications. Taking into account the advantageous and drawbacks of both network architectures, it becomes obvious that a combination of them is the logical consequence for future mobile radio networks. The scope of this paper is to introduce a future radio network architecture called hierarchical multihop cellular network (HMCN), which includes several multihop cells. The overlaying cellular network coordinates multihop cells in order to reach an optimal load balance in the whole network. The next evolutionary step towards a HMCN structure is to introduce multihop capable nodes (MHN), which can be fixed or even mobile. The paper shows the challenges and benefits of multihop with fixed nodes, especially the placement of fixed MHN and reduction of radio protocol overhead in order to provide high performance. When MHN are normal mobile nodes, multihop routing becomes a key point in the HMCN concept. The paper proposes a cellular based routing scheme and compares it with the ad hoc routing algorithms being developed by IETF.

This contribution highlights the possibilities and service characteristics for a cost and time efficient roll out for public wireless local area networks providing wide coverage in urban areas. The HIPERLAN/2 (H/2) multihop concept is based on intermediate relay stations. This eliminates the need for intensive cabling, reducing the costs. The H/2 multihop concept introduces a new element called forwarding mobile terminal (FMT) to the H/2 world. The FMT is a modified H/2 mobile terminal which only needs different software, but no additional transceiver. In city scenarios, the forwarding concept allows a fast roll-out while limiting the cabling cost and providing a reasonable user service in a large service area.

A concept is presented that introduces a new element called forwarding mobile terminal (FMT) to the HIPERLAN/2 (H/2) world. The FMT is designed to provide users in uncovered areas with access to the fixed Internet by providing an intermediate hop. The FMT is introduced in a way that it is backward compatible to existing H/2 equipment. The FMT is a modified H/2 mobile terminal which needs different software, but no additional transceiver. The concept is evolved for the H/2 link layers and signalling flows, furthermore future possible enhancements are presented.