Di Kong

Transport layer performance in 5G mmWave cellular

The millimeter wave (mmWave) bands are likely to play a significant role in next generation cellular systems due to the possibility of very high throughput thanks to the availability of massive bandwidth and high-dimensional antennas. Especially in Non-Line-of-Sight conditions, significant variations in the received RF power can occur as a result of the scattering from nearby building and terrain surfaces. Scattering objects come and go as the user moves through the local environment. At the higher end of the mmWave band, rough surface scatter generates cluster-based small-scale fading, where signal levels can vary by more than 20 dB over just a few wavelengths. This high level of channel variability may present significant challenges for congestion control. Using our recently developed end-to-end mmWave ns3-based framework, this paper presents the first performance evaluation of TCP congestion control in next-generation mmWave networks. Importantly, the framework can incorporate detailed models of the mmWave channel, beamforming and tracking algorithms, and builds on statistical channel models derived from real measurements in New York City, as well as detailed ray traces.

Impact of Low-Frequency Radar Interference on Digital Terrestrial Television

This paper considers low-frequency radar cohabitation with digital video broadcasting terrestrial (DVB-T). The spatial and temporal radio frequency propagation characteristics of the terrestrial and air-to-ground channels are modeled using state-of-the-art 3-D ray tracing for real-world broadcast television scenarios. Rooftop DVB-T receivers are assumed at user locations in two different urban environments, namely, the cities of London and Bristol, U.K.. The propagation characteristics are combined with appropriately orientated 3-D antenna radiation patterns, which include full polarization information for the DVB-T rooftop antennas, the base-station antennas, and the airborne radar antennas. A radar waveform generator is used to synthesize a representative range of radar signals. A DVB-T physical layer simulator has been developed to predict the performance of the received broadcast stream at each rooftop location. Factors, such as additive noise, the power and structure of the wanted DVB-T and the interfering radar signals, and the nature of the radio channels, are shown to influence the DVB-T bit error rate results. Statistics for the probability of high-quality DVB-T reception are reported as a function of the modulation scheme, propagation environment, aircraft orientation, radar antenna array configuration, and radar waveform type.

Throughput sensitivity to antenna pattern and orientation in 802.11n networks

In this paper the throughput and packet error rate for an in-home 802.11n network is theoretically derived for two different types of 3x3 antenna configurations. Our first configuration assumes the use of three low directivity omni-dectional elements. The second arrangement makes use of three orthogonally orientated directional elements. The spatial and temporal characteristics of the in-home channels are modelled using 3D ray tracing and combined with appropriately orientated complex polarmetric patterns for each antenna element. Physical layer throughput is computed for all modulation and coding schemes using a received bit information rate abstraction technique. The theory shows that directional antennas outperform the omni-directional devices in most cases. Directional elements show increased sensitivity to orientation, however for 83% of locations and orientations they still result in throughput enhancement. Directional antennas provide a 33% improvement in average data rate for random client orientations, improving to 52% with optimum alignment to the multipath.

Impact of BS antenna number and array geometry on single-user LTE-A data throughputs in realistic Macro and Pico cellular environments

This paper evaluates the theoretic performance of single-user multi-stream beamforming for heterogeneous LTE-A urban deployments (i.e. Macro and Pico cells). The work investigates the impact of six different Base Station (BS) antenna configurations on system level capacity. Our standard configuration assume 8 BS and 8 User Equipment (UE) antenna elements. To enhance capacity the BS antenna number was increased to 12 and 16. A 3D laser-scanned database for the city of Bristol (UK) was used as input to the 3D channel propagation model along with measured 3D complex voltage and polarimetric antenna patterns for the individual BS and UE antenna elements. More than 50,000 ray-traced Pico and Macro cellular links per BS configuration were investigated to ensure statistical relevance. Our analysis quantifies the capacity of an 8×8 Single-User Multiple-Input-Multiple-Output (SU-MIMO) solution in realistic urban heterogeneous environments. Results address the system level benefits of increasing the number of BS antenna elements as well as the sensitivity of capacity to vertical and horizontal spatial element configurations.

Closed-Loop Antenna Selection for Wireless LANs with Directional & Omni-Directional Elements

Throughput and packet error rate are analysed in a home environment for two different 3x3 wireless LAN solutions. A 3x3 EBF approach (using three radio chains) is compared with a reduced cost 2x2 architecture (using two radio chains). In the 2x2 solution the optimum antenna pair is selected from the same set of three antennas at the AP and client. The impact of directional, as well as omni- directional, antenna elements is considered. The spatial and temporal characteristics of the in-home channels are modelled using 3D ray tracing and combined with appropriately orientated complex polarmetric patterns for each antenna element. Physical layer throughput is calculated for all modulation and coding schemes and (for the 2x2 case) all possible antenna combinations using a novel received bit mutual information rate abstraction technique. Results show that antenna number, pattern and orientation all play a key role in determining the performance of an 802.11n system. As expected, 3x3 EBF outperforms the 2x2 solution; however, with optimum antenna selection the performance of 2x2 EBF is competitive, especially when directional antennas are used at low signal to noise ratios. For distant rooms, 3x3 EBF is only 15% better (in terms of throughput) than 2x2 EBF when directional antennas and dynamic antenna selection are applied. 2x2 EBF with omni antennas results in a 45% reduction in throughput (compared with 3x3 EBF).

Evaluation of 802.11 and LTE for Automotive Applications

This paper evaluates coverage prediction and radio performance for automotive infotainment applications. The work compares IEEE 802.11af operating in the TV White Space (TVWS) bands with LTE operating at 800MHz and 2.6GHz. Real world rural and urban databases are incorporated into a 3D ray tracing tool for the purposes of exploring signal coverage and data rates along with virtual drive tests. Results are presented as a function of radio standard, operating band, vehicle speed, user location and packet size. LTE is shown to outperform 802.11af since it is less affected by the multipath channel and the significant levels of Doppler Spread experienced by vehicular users. Results also indicate that 802.11af systems provide superior coverage, especially in rural environments.

A raptor enabled data carousel for enhanced file delivery and QoS in 802.11 multicast networks

802.11 WLANs do not provide any standardized solution for reliable data multicast. Multicast packets are delivered to multiple clients as a simple broadcast service without support for Automatic Repeat Request. Hence, a fixed low speed (robust) transmission mode is generally used to improve the reliability of multicast files. However, this results in the inefficient use of bandwidth. This paper details a reliable and efficient Wi-Fi multicast delivery solution for use in challenging outdoor environments. We propose an Application Layer Forward Error Correction enabled data carousel for reliable multicast transmission over standard 802.11 WLANs. To quantify the benefits of the proposed system, results are reported from a cross-layer simulator combining novel outdoor ray-tracing, a Physical layer abstraction simulator (to rapidly quantify the radio performance), a RaptorQ enabled multicast data carousel simulator and an optimal access point deployment tool. The simulation results demonstrate that RaptorQ enabled carousels (compared to standard carousels) significantly reduce the average response time and increase the percentage of satisfied users in a multicast network.

The Impact of Regulatory Transmit Power Constraints on the Relative Performances of Wi-Fi Beamforming and Antenna Selection

This paper analyses the theoretic throughput and coverage for a multi-element in-home Wireless Local Area Network (WLAN) in the 5 GHz band. In particular we study the impact of regulatory transmit power constraints on the relative performances of beamforming and antenna selection at the access point. A novel methodology is applied based on part measurement and part simulation. In home results are provided for standard 802.11n multiple-input multiple-output (MIMO) deployments. Results show that beamforming is particularly affected by regulatory transmit power constraints. For long-range links beamforming data throughputs are reduced by more than 50%. The average data rate in our test home reduces by 14.3% when regulatory constraints are applied to the beamforming system. For antenna selection the regulatory limits result in an 8.9% reduction in average throughput. Furthermore, we show that for long range links the average data rate is improved by 50% when the antenna selection is used in place of beamforming. Although antenna selection uses single stream MIMO more frequently than beamforming, the latter performs better when multiple spatial streams are used. Beamforming is shown to work best in the easy and medium-range channels.

Enhancing In-Home 802.11 Performance: Mesh or MIMO?

This paper analyses the throughput and coverage of an in-home wireless LAN in the 5GHz band. A novel approach is proposed based on part measurement, part simulation. The spatial and temporal characteristics of typical in-home channels are modelled using 3D ray tracing and combined with measured and appropriately orientated complex polarmetric patterns for each antenna element. Physical layer throughput is computed across 10 rooms in a virtual test house for all modulation and coding schemes. To accelerate the simulation process a received bit information rate abstraction technique is applied. Results are given for a standard 802.11a deployment. The benefits of 3x3 MIMO are then compared against the deployment of a Wireless Mesh Network (WMN). In the latter case meshing is performed using low cost single antenna radios. Relative to the benchmark SISO system, 3x3 MIMO is shown to improve the average throughput by 123%. Given four randomly deployed mesh points, the mesh solution improves the average throughput by 31%. Importantly, for the test house under consideration the MIMO solution improves connectivity by just 3%, while the mesh network offers 100% (whole home) connectivity. The mesh network (using 4 additional mesh points) is shown to support high definition video for 90% of the links, compared to 71% with 3x3 MIMO. Results confirm that mesh networks benefit the weaker links, while MIMO boosts the stronger links.

Evaluating the effect of antenna tilt and rotation on antenna performance in an indoor environment

In this paper the relative performances of antenna elements are assessed by combining the measured full (3D) far-field radiation patterns of each with measured angle of arrival data for a laboratory/office space. Three element types (Cavity-backed Slot, Printed Inverted-F and Dielectric Resonator) at two positions on a small mobile terminal are considered, and the terminal is tilted and rotated in order to determine the variation in performance that is obtained for three transmitter locations. Results are presented as cumulative distribution functions of directivity relative to a benchmark Hertzian dipole for operation at 5.2GHz. A variation in signal strength of greater than 30dB was observed due to both the directivity of the elements and their polarisation alignment with the vertical transmitting source.