With balanced system performance, implementation complexity and hardware cost, hybrid antenna array is regarded as an enabling technology for massive multiple-input and multiple-output communication systems in millimeter wave (mmWave) frequencies. Angle-of-arrival (AoA) estimation using a localized hybrid array faces the challenges of the phase ambiguity problem due to its localized nature of array structure and susceptibility to noises. This paper discusses AoA estimation in an mmWave system employing dual-polarized antennas. We propose an enhanced AoA estimation algorithm using a localized hybrid dual-polarized array for a polarized mmWave signal. First, the use of dual-polarized arrays effectively strengthens the calibration of differential signals and resulting signal-to-noise ratio with coherent polarization combining, leading to an enhanced estimate of the phase offset between adjacent subarrays. Second, given the phase offset, an initial AoA estimate can be obtained, which is used to update the phase offset. By employing the updated one, the AoA is re- estimated with improved accuracy. The closed-form mean square error (MSE) lower bounds of AoA estimation are derived and compared with simulated MSEs. The simulation results show that the proposed algorithm in combination with hybrid dual- polarized arrays significantly improves the estimation accuracy compared with the state of the art.

In this paper, we study the low complexity reception of multiuser signals in uplink millimeter wave (mmWave) communications using a partially connected hybrid antenna array. Exploiting the mmWave channel property, we propose a low-complexity user-directed multiuser receiver with three novel schemes for allocating subarrays to users. This receiver only requires the knowledge of angles-of-arrival (AoAs) for dominating paths and a small amount of equivalent channel information instead of perfect channel state information. For comparison, we also derive a successive interference cancellation-based solution as a performance benchmark. We design two types of reference signals with the channel estimation method to enable efficient and simple estimation for AoA and equivalent baseband channel. Also, we provide analytical results for the performance of the AoA estimation, using the lower bounds of mean square errors in line-of-sight dominated mmWave channels. The simulation results validate that the proposed channel estimation method is effective when employed in combination with a zero-forcing equalizer.

Due to the limited dynamic range of the off-the-shelf electrical and optical components, deliberate digital clipping (DDC) is widely applied to optical orthogonal frequency division multiplexing (OFDM) based visible light communication systems. In this paper, we present a theoretical characterization of the layered asymmetrically clipped optical OFDM (ACO-OFDM) signals subject to peak clipping. We decouple a clipped L -layer ACO-OFDM symbol to L single-layer ACO-OFDM symbols, each corresponding to a layer, and show that these symbols are subject to symmetrical peak clippings at random levels. Using Bussgang's theorem, the resulting attenuation factors and variances of the additive noise associated with each layer are derived. It is shown that the clipping noise caused by the DDC mainly falls onto the first layer, and its impact is gradually reduced in the subsequent layers. In order to combat the clipping noise, a novel receiver based on decision aided reconstruction is proposed. Simulation results show that the proposed receiver can effectively mitigate the clipping noise, leading to significant improvement of bit error rates over the conventional receiver.

The superiority of exploring millimeter wave (mmWave) frequencies for future wireless communication systems has pushed forward the development of large-scale antenna arrays for achieving sufficient array gain and high spectral efficiency. In this paper, we study the matrix normalization (MN) based zero-forcing (ZF) hybrid precoding in multi-user multi-input-multi-output (MU-MIMO) mmWave systems. We derive the upper bounds of the achievable rate for two representative hybrid array structures, i.e., fully-connected structure and partially-connected structure. Analytical and simulation results validate the tightness of the proposed performance upper bounds for both hybrid structures using massive array, and provide a comparison of the achievable rate using MN and vector normalization (VN).

In this paper, fractional reverse polarity optical orthogonal frequency division multiplexing (FRPO-OFDM) is studied to enable dimming compatible visible light communications. The scheme combines a layered asymmetrically clipped optical OFDM (ACO-OFDM) sequence with an information-carrying brightness control sequence (BCS) in the form of M-ary pulse position modulation. We derive the expressions of the FRPO-OFDM signal and its achievable brightness level, and develop an effective detector which can recover information from both sequences based on maximum likelihood detection. We show that when the detector is to be implemented, the use of multi-layer ACO-OFDM imposes strong periodicity on the BCS, which leads to a trade-off between spectral efficiency and brightness resolution for dimming control. It is shown that high spectral efficiency can be achieved with practical dimming requirements. Simulation results show that the extra information carried by the BCS can be decoded with extremely low bit error rate and thus has negligible impacts on the demodulation of the ACO-OFDM signal, when the system nonlinearity is not dominating.

Visible light positioning (VLP) for indoor applications has recently received considerable attention. In this paper, an aperture based receiver is considered and the position is estimated based on the angle-of-arrival (AOA), because the AOA can be determined by comparing the relative differences in the received signal strengths (RSS) in the different photodiodes (PDs) of the receiver. Hence, in contrast to RSS based positioning solutions, knowledge of the transmitted signal strengths is not required. The performance of the proposed approach depends on the transmitted signal strength only through the dependency on the signal-to-noise ratio. To evaluate the accuracy of the estimates of the AOA, we derive the Cramer-Rao lower bound, and show that the expected error on the AOA is very small. Applying triangulation to determine the position of the receiver, it follows that centimetre accuracy can be obtained with the aperture-based receiver.

Due to the enormous needs for signal processing and hardware constraints, the full digital implementation for a large antenna array at mmWave frequencies becomes intractable. Hence, receiver design for MIMO system with massive hybrid array is very demanding, particularly for subarray structure. In this paper, we propose two low- complexity uplink multiuser receiver design schemes (single-beam and multi-beam) under the circumstances of only users' angles of arrival (AOAs) available at base station, which greatly simplify the analog beamforming structure and reduce the complexity of channel estimation especially when the number of antenna elements is considerably large. In the single-beam scheme, subarrays are mapped to different users one by one such that each subarray serves a specific user. In the multibeam scheme, all subarrays use the same analog beamforming and any subarray's beamforming signal potentially points at all users. Digital beamforming is then employed to combine all subarray signals followed by equalization. Simulations are performed to compare the proposed schemes with other schemes that need channel information available at the subarray output. Since only AOAs are required for analog and digital beamformings, the proposed schemes are more suitable for mmWave MIMO system with massive hybrid array in spite of slight sum-rate loss.

In this paper, we study a novel optical orthogonal frequency division multiplexing (OFDM) scheme, fractional reversed polarity optical OFDM (FRPO- OFDM), which enables dimmable visible light communications. The scheme combines an asymmetrically clipped optical OFDM (ACO-OFDM) sequence with an information-carrying brightness control sequence (BCS). We show that the new scheme can create a wide range of brightness levels whilst transmitting extra information using the BCS to enhance the spectral efficiency. The detector which can recover information from both sequences is derived based on maximum likelihood (ML) detection and estimation. It is shown that the information carried by the BCS using variable on-off keying (VOOK) can be detected with extremely low bit error rate (BER) and that the resulting overall BER of FRPO-OFDM is constant over a wide dimming range.

Spatial multiplexing (SMP) and optical spatial modulation (OSM) are two important technologies for indoor multiple-input multiple-output (MIMO) visible light communications (VLC). In this paper, we compare the performance of SMP and OSM systems which are operated with typical configurations in an indoor scenario. Both systems transmit data using asymmetrically clipped optical orthogonal frequency division multiplexing (ACO-OFDM) as the modulation scheme. It is shown that in order to achieve the same data rate with the SMP, the OSM has to employ large constellations which become impractical when either the number of luminaires or the constellation size of the SMP is greater than four. Simulation results are presented for both MIMO systems using four luminaires as transmitters and a receiver configured with three different front-ends. These receivers include a conventional non-imaging receiver, a prism-based receiver and an aperture-based receiver. The BER results demonstrate that SMP outperforms OSM in terms of both the size of the region in which a receiver can achieve low BER and the BER at typical receiver positions.

In this paper, we investigate the problem of indoor positioning using visible light systems. The directional detector array we use is comprised of a number of receiving elements, each consisting of an aperture and a photo diode, which are arranged to offer good angular diversity, and can be implemented within a compact receiver structure [1]. The receiving elements receive the light from a number of white LEDs, which are typically attached to the ceiling, and which act as anchors. In order to get an indication of the received signal strengths of the different LEDs, we average the received signals over time. The relative signal strengths in the different receiving elements do not only provide information on the distance between the LEDs and the detector array, but also about the angle-of-arrival of the light. By combining the information of the receiving elements, the position of the detector can be estimated. In order to assess the accuracy of positioning algorithms based on this approach, we derive the Cramer-Rao lower bound on the position accuracy. Assuming the white LEDs transmit an optical power of 1 W, and the time averaging is done over 1 millisecond, an accuracy of the order of a centimetre can be achieved.

In this paper, we present a new type of receiver for MIMO optical wireless communications. The new receiver consists of multiple receiving elements (REs), each made up of a photodetector (PD) placed under an aperture. The directionality of each RE is shown to depend on the relative position of the PD and the aperture. Receiving patterns are calculated for various MIMO receiver designs, demonstrating that significant angular diversity can be achieved within a compact receiver structure. Simulation results are presented for a number of typical indoor scenarios when the transmitted signals are modulated using asymmetrically clipped optical OFDM (ACO-OFDM). For the scenarios considered, the bit error rate is lowest in the middle of the room and highest in the corners. The results indicate that for typical parameter values, reliable communication can be achieved throughout the room.

This paper describes a new form of receiver for multiple-input multiple-output (MIMO) optical wireless communications. The new receiver uses a number of photodetectors (PDs), all facing in the same direction but with different fields of view (FOVs). This enables a compact planar receiver structure. The performance of the receiver is analyzed for an indoor visible light communications system where LED lights are used as data transmitters. It is shown that the MIMO channel matrix has full rank for typical receiver positions and that for a given number of PDs, the new receiver provides much greater diversity than a conventional receiver of the same overall dimensions in which all of the PDs have the same FOV. Finally the overall bit error rate (BER) is presented for systems using asymmetrically clipped optical OFDM (ACO-OFDM) as the modulation scheme. Results are presented for both zero forcing (ZF) and minimum mean square error (MMSE) equalization in the receiver. It is shown that the BER varies with receiver position, with relatively high values in the center and the corners of the room.

In this paper, we analyze the performance of a prism-array receiver in a multiple-input multiple-output (MIMO) intensity-modulated direct-detection optical wireless communication system. The elements of the MIMO channel matrix depend on the geometry of each prism and the orientation of the prism relative to the optical transmitter. An expression is derived for each channel gain as a function of the angle of the prism, the aspect ratio and area of the receiving surface of the prism, and the orientation and distance between the transmitter and receiver. Results are presented for the directionality and overall field of view (FOV) of a receiver comprising six prisms, showing that compact receivers can be designed, which combine well-conditioned channel matrices with wide FOVs. Simulation results are presented for a typical indoor visible light communication system where four light-emitting diode lights transmit information using asymmetrically clipped optical orthogonal frequency division multiplexing and linear equalizers are used in the receiver. It is shown that the overall bit error rate depends very strongly on the angle of the prisms and on the position of the receiver, but that a simple six-prism receiver can successfully decouple the four transmitted signals at typical receiver positions within the room.