Optical Wireless Communications (OWC) is one of many new and growing interests of the research community; however, a significant amount of system-level OWC research has focused on theoretical and/or simulated analysis with limited experimental validation. In order to address this concern and support experimental OWC test systems, we have previously introduced the gr-owc out-of-tree module in GNURadio, along with a detailed description of potential hardware configurations to integrate front-end OWC hardware with USRP SDR equipment. Our initial work also demonstrated the benefits of modular design in an SDR-based OWC test system where signal processing parameters can be configured for the various characteristics of different front-end OWC hardware. This prior work demonstrated point-to-point and multi-node OWC test systems, but the flowgraph parameters were mostly fixed at the time of execution.
In this article, we will introduce recent improvements that build upon our previous successes with gr-owc. Namely, we implement dynamic control and coordination of distributed OWC nodes from a centralized testbed controller. This is achieved through a combination of GNURadio’s XML-RPC and ZMQ modules to remotely modify parameters on running flowgraphs, along with bash scripts and the Secure Shell (SSH) protocol to remotely start/stop GNURadio flowgraphs at different nodes. This results in an OWC testbed architecture that is adaptable, versatile, and scalable.
To exemplify the testbed’s capabilities, we have demonstrated the application of centrally controlled subcarrier allocation for a multi-cell and multi-user OWC system implementing DC-Biased Optical OFDMA. For the purpose of performance analysis and comparison, the implementation of a centralized testbed controller has allowed for improvements in automated data collection in GNURadio. In particular, we can now implement automated packet error rate analysis for multiple scenarios (e.g., subcarrier assignments) with multiple data points for each scenario – ultimately reducing the person hours required to collect quantitative data at scale.
Beyond the benefits of automated data collection, this testbed architecture also enables many future investigations for dynamic parameter adaptation. In order to accommodate mobile devices with variable traffic load, practical multi-cell/multi-user OWC systems should be able to dynamically allocate resources amongst OWC cells and users. Our testbed architecture enables future demonstration and analysis of different techniques for interference mitigation, rate adaptation based on user demand, or inter-cell handover for mobile users. Insights gained from this experimental analysis could improve network performance in practical multi-cell/multi-user OWC systems and offer experimental validation of systems-level research existing in the literature.
In summary, our open-source testbed supports the investigation of new resource allocation techniques and iterative improvements to the baseline OWC techniques that we provide within gr-owc, fostering collaborative research and pushing the capabilities of future OWC systems. In our paper we will provide a detailed description of our testbed’s implementation and requirements for deployment, and we will highlight the potential capabilities that our testbed can enable.
|Talk Length||15 Minutes|
|Link to Open Source Code||https://github.com/UCaNLabUMB/gr-owc|