Patrick Bidigare

Downlink Training Techniques for FDD Massive MIMO Systems: Open-Loop and Closed-Loop Training With Memory

The concept of deploying a large number of antennas at the base station, often called massive multiple-input multiple-output (MIMO), has drawn considerable interest because of its potential ability to revolutionize current wireless communication systems. Most literature on massive MIMO systems assumes time division duplexing (TDD), although frequency division duplexing (FDD) dominates current cellular systems. Due to the large number of transmit antennas at the base station, currently standardized approaches would require a large percentage of the precious downlink and uplink resources in FDD massive MIMO be used for training signal transmissions and channel state information (CSI) feedback. To reduce the overhead of the downlink training phase, we propose practical open-loop and closed-loop training frameworks in this paper. We assume the base station and the user share a common set of training signals in advance. In open-loop training, the base station transmits training signals in a round-robin manner, and the user successively estimates the current channel using long-term channel statistics such as temporal and spatial correlations and previous channel estimates. In closed-loop training, the user feeds back the best training signal to be sent in the future based on channel prediction and the previously received training signals. With a small amount of feedback from the user to the base station, closed-loop training offers better performance in the data communication phase, especially when the signal-to-noise ratio is low, the number of transmit antennas is large, or prior channel estimates are not accurate at the beginning of the communication setup, all of which would be mostly beneficial for massive MIMO systems.

Receiver-coordinated distributed transmit beamforming with kinematic tracking

A distributed transmit beamforming technique is described for a scenario with two or more transmit nodes and one intended receiver. The protocol includes a measurement epoch, feedback from the intended receiver to the transmit nodes, and a beamforming epoch. The intended receiver tracks the clock and kinematic parameters of the independent transmit nodes and coordinates the transmit nodes by feeding back state predictions which are then used as phase corrections to facilitate passband phase and frequency alignment at the receiver. A three-state dynamic model is developed to describe the stochastic kinematics and clock evolution of each transmit node relative to the frame of the receiver/coordinator. Steady-state analysis techniques are used to analytically predict the tracking performance as well as the beamforming gain as a function of the system parameters. Numerical results show that near-ideal beamforming performance can be achieved if the period between successive observations at the receiver/coordinator is sufficiently small.

Receiver-coordinated distributed transmit nullforming with channel state uncertainty

A distributed coherent transmission scheme in which two or more transmit nodes form a beam toward an intended receiver while directing nulls at a number of other “protected” receivers is considered. Unlike pure distributed beamforming, where the ith transmit coefficient depends only on the ith transmit nodes channel to the intended receiver, the transmit coefficients of a distributed nullformer each depend on the channel responses from all of the transmit nodes to all of the protected nodes. The requirement for each transmit node to know all of the channels in the system makes distributed transmit nullforming challenging to implement in the presence of channel time variations. This paper describes a receiver-coordinated distributed transmission protocol, in the context of a state-space dynamic channel model, in which the receive nodes feed back periodic channel measurements to the transmit cluster. The transmit nodes use this feedback to generate optimal channel predictions and then calculate a time-varying transmit vector that minimizes the average total power at the protected receivers while satisfying an average power constraint at the intended receiver during distributed transmission. We demonstrate via analysis and numerical simulation the efficacy of the technique even with low channel measurement overhead, infrequent update intervals, and significant feedback latency.

Receiver-coordinated zero-forcing distributed transmit nullforming

A coherent cooperative communication system is proposed in which a distributed array of transmit nodes forms a beam at a desired receiver while simultaneously steering nulls at several protected receivers. Coherent transmission is achieved through a receiver-coordinated protocol where the receivers in the system use state-space channel tracking and provide feedback to the transmit cluster to facilitate distributed transmission. Analytical estimates for the performance degradation in the nulls due to channel estimation errors are verified by simulations. Numerical results demonstrate that the technique is effective even with low channel measurement overhead, infrequent measurement intervals, and feedback latency.

Implementation and demonstration of receiver-coordinated distributed transmit beamforming across an ad-hoc radio network

Distributed transmit beamforming using an ad-hoc network of 10 RF transmitters was demonstrated using radio nodes developed from off-the-shelf components and modules. A time-slotted protocol allowed carrier phases from each transmitter to be measured at a receiver and fed back to the transmitters where Kalman filters were used to predict the offset phases and frequencies. Offsets were digitally compensated for during beamforming intervals. Beamforming gain within 0.1dB of ideal was demonstrated across 1 km at 910MHz. This is the first report (to our knowledge) of a successful outdoor RF distributed transmit beamforming experiment using independent clocks at this scale.

A closed-loop training approach for massive MIMO beamforming systems

There has been a growing interest in wireless systems that employ a very large number of transmit antennas. Some theoretical results have shown that substantial improvements in network capacity are possible. Despite this work, a major challenge is how these large transmit arrays should perform training to allow receiver channel estimation. Without new techniques, the heavy burden of training could overwhelm the system and mitigate any possible improvements, especially in systems using frequency division duplexing (FDD) where channel reciprocity cannot be exploited. In this work, we propose the use of closed-loop training. In this framework, the transmitted training signal is optimized to improve data communications performance by using prior information about the current channel obtained from past channel estimates. The work focuses on block-fading channels with temporal and spatial correlation. Simulation results show improved performance.

DSP-centric algorithms for distributed transmit beamforming

Distributed transmit beamforming is a means of increasing range and power efficiency via local collaboration among neighboring nodes in order to transmit a common message to a remote destination. While its basic feasibility has been established by recent analyses and prototypes, transitioning this concept to applications requires the development of protocols and architectures which can be implemented efficiently using digital signal processing (DSP). In this paper, we describe DSP-centric algorithms and their performance limits, and report on recent results from simulations and software-defined radio experiments.

Attaining fundamental bounds on timing synchronization

In this paper, we propose an algorithm for timing synchronization that attains fundamental bounds derived by Weiss and Weinstein. These bounds state that, in addition to improving with time-bandwidth product and signal-to-noise ratio (SNR), timing accuracy also improves as the carrier frequency gets larger, if the SNR is above a threshold. Our algorithm essentially follows the logic of the Weiss-Weinstein bound, and has the following stages: coarse estimation using time domain samples, fine-grained estimation using a Newton algorithm in the frequency domain, and final refinement to within a small fraction of a carrier cycle. While the results here are of fundamental interest, we are motivated to push the limits of synchronization to enable the tight coordination required for emulating virtual antenna arrays using a collection of cooperating nodes.

Statistical Modeling and ML Parameter Estimation of Complex SAR Imagery

Accurate statistical models for the complex pixels forming fine-resolution synthetic aperture radar (SAR) images are needed for several engineering applications, including coherent signal detection in SAR clutter, automatic target recognition, and automatic SAR RCS calibration without calibration targets. We derive the maximum likelihood estimator for the parameters of a complex generalized Gaussian distribution and show that it can be efficiently computed. Applying this to fine-resolution SAR images representing a wide variety of scene contents, we show that this model very accurately captures both the central regions and tails of the data distribution.

Scalable feedback algorithms for distributed transmit beamforming in wireless networks

We explore a class of techniques for distributed transmit beamforming where the beamforming target sends cumulative feedback that is broadcast to all of the beamforming nodes. The simplest technique in this class is a 1-bit RSS feedback algorithm that has been studied in detail in the literature. Under this 1-bit algorithm, transmitters make random phase perturbations and the receiver periodically sends 1 bit of feedback indicating whether the received signal strength has increased or not compared to what was observed in the past. While this simple algorithm has very attractive properties such as dynamic tracking of time-varying phases, robustness to noise and other disturbances and is also simple to implement, we show in this paper that it also has serious limitations such as slow convergence and poor tracking performance in the presence of frequency offsets between the transmitters. We then show that enhanced feedback algorithms where the receiver sends as feedback several bits of feedback indicating the amplitude and phase of the received signal over time, are able to achieve beamforming in the presence of frequency offsets and large feedback channel latencies, while retaining the scalability and robustness of the 1-bit algorithm.