This work investigates the capacity-approaching input distribution (CAID) of the multiple-input single-output (MISO) optical intensity channel (OIC) in the presence of both per-antenna peak- and average-power constraints. Although OICs have been extensively studied in the literature, prior studies predominantly focus on capacity bounds and asymptotic behaviors, with limited attention directed towards CAIDs of OICs, particularly in the context of multi-input systems. For the first time, the exact CAID of the MISO OIC is determined by a successive minimum-power mapping (SMPM) strategy, in which the CAID problem is decomposed as an upper-level power allocation subproblem and a series of lower-level optimal input probability distribution (OIPD) subproblems for single-input single-output (SISO) cases. Adopting a hierarchical iteration structure, the binary search and successive linear least square algorithms are designed to address the power allocation subproblem in the outer iteration, while the layered primal-dual interior-point algorithm is employed to solve the SISO-OIPD subproblem in the inner iteration. The derived CAIDs for MISO OICs are presented and analyzed, and the corresponding mutual information results are compared with several theoretical bounds to validate the asymptotic tightness. It has been shown that the proposed CAID is structurally suboptimal but numerically approaches the theoretical capacity, especially in both low and high SNR regimes.

Inter-user interference (IUI) and inter-cell interference (ICI) are critical challenges faced by dense deployments of cellular visible light communication (VLC) systems, especially in scenarios involving Internet of Things (IoT) devices with quality of service (QoS) requirements. Thus, advanced techniques, such as rate splitting multiple access (RSMA), have been explored to deal with these challenges. However, the existing solutions based on RSMA primarily focus on spectral efficiency while neglecting the issue of fairness, i.e., the minimum achievable rate among users. Consequently, we propose a novel coordinated multipoint (CoMP) assisted cooperative RSMA scheme for aggregated VLC/radio frequency (RF) cellular systems to simultaneously address these issues. To maximize spectral efficiency under QoS constraints and enhance fairness, we formulate optimization problems to jointly optimize precoding and rate allocation. Given the inherent non-convexity of these problems, we apply successive convex approximation (SCA) and semidefinite relaxation (SDR) techniques to iteratively and effectively obtain near-optimal solutions. Experimental results demonstrate that, compared to the baseline schemes, the proposed scheme exhibits superior performance across varying performance metrics, user deployments and network loads.

With the widespread application of light emitting diodes (LEDs) and the rapid evolution of visible light communication (VLC), visible light-based positioning (VLP) is envisioned as a promising solution for indoor positioning due to its high accuracy and environmental friendliness. However, the performance of VLP is significantly affected by the unknown orientation of the user equipment (UE) and system parameters uncertainty. Besides, high computation complexity and vulnerability to hardware degradation of the VLP algorithms are also the challenges to meet. To handle the above issues, a novel cone geometry-based simultaneous position and orientation estimation (SPO) scheme with low complexity is proposed. By modeling the field of view (FOV) of the photodiode (PD) as a cone, the geometry relationship between the sensed LEDs and the UE position and orientation is established. The binary values of whether the LEDs signals are received or not are exploited, and an alternating flashing strategy for the LED arrays is designed to guarantee normal illumination and reduce bandwidth consumption. Moreover, the theoretical lower bound derived from Kanatani-Cramer-Rao bound (KCRB) is provided to analyze optimality of the estimator. Simulation results demonstrate the proposed scheme achieve an excellent performance gain with low complexity over two classical SPO baseline, and the critical factors that affect the performance are investigated.

In this paper, a downlink non-orthogonal multiple access (NOMA) system with partial channel state information (CSI) in mobile scenarios is considered. In this case, the channel gain of each user will vary as its position changes, which has an influence on spectrum efficiency of conventional NOMA. An analytical framework is developed to evaluate the impact of estimated position deviation in terms of decoding order error probability and average sum rate performance. To further improve the system performance, a novel NOMA scheme based on position filtering is proposed. Monte Carlo simulation is also provided to demonstrate the improvement of spectrum efficiency compared with conventional NOMA and OMA schemes. And the results verify the accuracy of proposed analytical framework.

We propose a novel robust H.263+ video transmission scheme using a partial backward decodable bit stream (PBDBS), which reverses the bit stream of some coded macroblocks (MBs) so that these coded MBs can be decoded in a backward direction. Because of using variable-length coding (VLC), the baseline H.263+ coder will lose synchronization until the next resynchronization marker when errors occur in the coded bit stream. It is wasteful to discard all of the remaining coded MBs in a group of blocks (GOB). Our method is to alleviate this effect by reversing the bitstream that belongs to the coded second half row of MBs in a GOB. When a decoder loses resynchronization, it will search the next resynchronization marker and decode the bitstream backward. It can retrieve half row of MBs that are lost otherwise. This method combined with error concealment can greatly improve the transmitted image quality. Up to 2-dB luminance peak signal-to-noise ratio improvement can be found in our method compared with the baseline H.263+ coder. We also compare our PBDBS method with H.263++ Annex V using data partitioning and reversible VLC. Under lower encoder and decoder complexity, the PBDBS method obtains more pleasant results than Annex V, and it can be utilized in noisy channels such as wireless networks.