Multi-carrier modulations over sparse channels: communication, channel estimation, and radar sensing

Abstract

Every year, in every field, the research introduces many novelties to contribute to the improvement of the state-of-the-art technologies. Under this context, the upgrade in terms of hardware capabilities and functionalities necessitates a constantly updated definition of the software, to be able to exploit all the possibilities owned by each single equipment. By mostly focusing on communication, i.e., the capability to send useful information from a source to a destination, the research looks forward to the definition of new communication standards, to break the limits imposed by outdated mechanisms and rules. Under this context, the upcoming technology for wireless and digital communications is the 5G standard, which brings improvements in both hardware equipment and software definition. Differently from its predecessor, i.e., the LTE (4G) standard, many new use cases and scenarios will be introduced. For instance, the improvements move towards the maximization of the data rate, to support the growing and massive amount of traffic data (especially from mobile users), the minimization of the latency, to enable new delay-intolerant services, and to optimize the energy consumption of any device, together with its cost. From a more technical point of view and of relevant interest for this dissertation, new vehicular communication (V2V, V2X) scenarios are going to be studied, by mostly focusing on millimeter wave frequency bands, strictly linked to the definition of massive MIMO antenna systems exploiting huge antenna gains through beamforming, necessary to contrast the high propagation loss typical of that frequencies. Under this context, against common and well known radar systems, able to efficiently detect and locate a target within the range-velocity plane, new joint radar and communications techniques are taking part of the current literature. These system are mainly focused on the transmission of useful information towards the targets, which are, thus, not only “passively” detected, and a single equipment is able to perform both operational modes, avoiding to split the functionalities between two distinct subsystems (with increased cost and complexity). Hence, there are mainly two approaches to solve the aforementioned problem. The first one considers the application of common radar waveforms, adapted to carry useful information with them. The second one, which is the one explored in this dissertation, takes into account typical communication waveforms (single- or multi-carrier), and, while communication tasks come naturally, the radar processing is performed with novel methods exploiting the knowledge of the transmitted information (known by both transmitter and receiver, if physically colocated), thus differs from a more direct, or “radar-like”, threshold analysis of the backscattered power from the target. The choice of the communication waveform is subject to a non trivial tradeoff. On one hand, the system aims the maximization of the communication achievable rate, i.e., the amount of information sent in a time-frequency window. On the other hand, radar tasks have to be performed with as much precision as possible, in order to correctly localize a target in all dimensions, i.e., range, velocity, and space (angular) location. Typically, pure radar tasks are performed with chirp-like pulses, i.e., short single-carrier impulses with large bandwidth, such that the total energy delivered towards the targets is compensated by the band occupation of the signal. Thus, the joint definition of a short pulse, together with large bandwidth, leads to a very precise localization of the target over the three aforementioned domains. However, the amount of (possible) useful information, impressed on top of such chirp, is poor. A solution to improve the communication rate is the use of multi-carrier digital waveforms, modulating information symbols not only in time domain (as single-carrier) but also in the frequency band, split in many subcarriers each occupied by a different modulation symbol. However, limitations are linked to the definition of the symbol time and the subcarrier spacing, which are in a one-to-one relation, and a good localization is not only challenging, for instance, in terms of signal processing algorithms, but also definitely sub-optimal with respect to single-carrier solutions, but this is the cost to pay in order to bring communication features together with radar tasks. In conclusion, the current literature is moving towards the definition of new multi-carrier schemes able to break the limits, in terms on communica- tion rate, imposed by classical radar waveforms, and the optimization of the tradeoff between the two different tasks is an open problem, whose optimal solutions have not been defined yet. The choice of the multi-carrier modulation for joint radar and communica tion falls into two distinct waveforms, i.e., orthogonal frequency-division mul- tiplexing (OFDM) modulation and orthogonal time frequency space (OTFS) modulation. OFDM is the most popular multi-carrier modulation of recent years, widely studied and standardized in most of the current communication standards, including 5G. The motivation of this choice is simple: thanks to the application of a cyclic prefix between symbols, i.e., a guard interval to prevent inter-symbol interference, and under the assumption of absence of inter-carrier interference, which holds under reasonable amount of the Doppler effect and subcarrier spacing, the communication channel can be diagonalized and symbol-by-symbol detection performed. Clearly, the appealing simplicity of detection makes OFDM the best choice for modern digital communications. On the other hand, OTFS is a modulation waveform with two big differences with respect to its direct competitor. First, it does not necessitate the insertion of the cyclic prefix, achieving a better communications rate, i.e., more information is sent over a time-frequency window, but at the cost of a more complex detection approach, working blockwise and not symbol-by-symbol. Second, OTFS is not sensitive to delay and Doppler shifts, meaning that its performance is kept constant whatever the distance and the speed between transmitter and (target) receiver. This feature is very appealing for joint radar and communication tasks, being the scenario very dynamical, with possible remarkable Doppler shifts and delays, increased by considering the round-trip time between radar transmitter and target. Based on the aforementioned analysis, in this dissertation we take care of a fair comparison between the two digital modulation formats, from the point of view of radar, parameter estimation, achievable communication rate, channel estimation, and more other tasks, to determine their positive and negative aspects, such that a system designer is able to choose the most suitable waveform for a given scenario or application.