In wireless communications systems, signals can be transmitted as time (temporal) or spatial variants across 3D space, and in both ways. However, using temporal variant communication channels in high-speed data transmission introduces inter-symbol interference (ISI) which makes the systems unreliable. On the other hand, spatial diversity in signal processing reduces the ISI and improves the system throughput or performance by allowing more signals from different spatial locations at the same time. Therefore, the spatial features or properties of visible light signals can be very useful in designing a reliable visible light communication (VLC) system with higher system throughput and making it more robust against ambient noise and interference. By allowing only the signals of interest, spatial separability in VLC can minimize the noise to a greater extent to improve signal-to-noise ratio (SNR) which can ensure higher data rates (in the order of Gbps-Tbps) in VLC. So, designing a VLC system with spatial diversity is an exciting area to explore and might set the foundation for future VLC system architectures and enable different VLC based applications such as vehicular VLC, multi-VLC, localization, and detection using VLC, etc. This thesis work is motivated by the fundamental challenges in reusing spatial information in VLC systems to increase the system throughput or gain through novel system designing and their prototype implementations.