Using Overhead Lights for Wireless Communications Promises to Substantially Increase the Capacity of Existing Wireless Networks
Outcome/Accomplishment
Researchers from the NSF Smart Lighting Center have demonstrated the capacity for novel energy-efficient LED lighting to provide data access networks that are complementary to existing wireless systems. By modeling the line-of-sight characteristics of light and the capacity for light to be used to communicate data bits, LED lighting has been shown to have very high capacity for supporting wireless communications that have now become ubiquitous. Moreover, light-based wireless communications can ‘reuse’ the available lighting spectrum in very small spaces – small lighting cells that are placed at approximately 1m apart can serve independent wireless channels whereas existing wireless (WiFi) transmitters compete for available spectrum. Center researchers demonstrated how the combination of the LED light-based cells and existing WiFi can complement each other in providing wireless access that can scale – expand to meet increasing data demand, by the addition of lighting cells and/or WiFi access points.
Impacts and benefits
With the evolution of mobile telephony and personal computing to modern smartphones and tablets has come a dramatic increase demand for wireless data delivery. These mobile devices are readily available to render rich media delivery including video and their users are actively generating this data traffic demand, sometimes beyond the limits of the network service providers. Cisco forecasts that fixed position wireless traffic will see a 39% annual growth rate between 2010 and 2015; however the capacity of RF communication techniques has not seen comparable gains in recent years. While user demand is increasing much faster than RF capacity, we have begun to see a wireless traffic jam in the RF spectrum. Without mitigation, we are on a path of ‘data starvation’ as mobile platforms are increasingly adopted in existing applications and in emerging applications including heath care, transportation, and commerce.
The NSF Smart Lighting ERC team approached this challenge from a systems perspective – by understanding the multiple interacting demands imposed on indoor environments related to health, safety, productivity, and energy efficiency and by leveraging the skills of a diverse team of scientists and engineers in the context of lighting systems. The result is a technical solution that intersects critical societal outcomes with diverse specialized knowledge for constructing energy efficient LED-based communications. The critical discovery reported here establishes the motivation for the implementation of a cooperative system that uses WiFi and LED lighting devices in development to provide scalable performance and operating characteristics relating to wireless communication.
Explanation and Background
The work stems from the need for additional wireless capacity due to growing demand for wireless services and applications. Two trends that come from CISCO’s Visual Network Index report show that consumer traffic will be dominated by internet video and that the majority of wireless traffic is expected to come from “fixed-position” wireless devices. Considering both of these trends, cooperative systems as proposed in this work provide enormous benefits for wireless communication. First, the asymmetric nature of the proposed system is a great fit for internet video systems which are predominantly downstream traffic. This alludes to the idea that a large percentage of wireless traffic caused by video streaming can be offloaded from the RF medium to the Visible Light Communication (VLC) channels. Additionally, the indication that most wireless traffic comes from fixed-position devices would allow a mobile user to move to a VLC hotspot and remain in place during use – hence removing much of the overhead associated with handover as a user traverses the environment.
In a cooperative system, VLC capable users benefit from higher data rates and channel reuse provided by VLC cells while the removal of congestive traffic from the RF medium provides benefits to non-VLC users. As users contend for the RF medium, the probability of a packet collision is reduced as high data rate video streams are offloaded to the VLC channels. This improves the throughput of the RF channel while adding the capacity provided by VLC channels – leading to drastic gains in aggregate throughput within the environment. We have shown through simulation that these gains are scalable with the number of VLC channels and that a cooperative system performs better than either system acting alone under static conditions.
Current research within the Smart Lighting ERC is focusing on optimal protocols for traffic distribution with considerations for mobile users and dynamic signal conditions. As a user traverses through an environment, they should dynamically connect to the network via the optimal channel. The concept of handover determines when the traffic flow from AP to user should be routed to a different channel. This can include transfer from one cell to another or between the RF and VLC channels. Simple protocols observe received signal from multiple channels and select the optimal; however the research in the ERC focuses on predictive methods that account for estimated future conditions in the handover decision. Since there is innate overhead involved in handover, predictive methods allow a system to weigh the benefit of a handover versus the overhead necessary to actually reroute traffic. When considering a mobile user, predictive handover decisions estimate a user’s motion path in order to determine when a handover is unnecessary – as in the case when a user is passing through the outer edge of a cell. Additionally, they can observe VLC signal conditions to predict whether a signal loss is due to a blocking condition or a user moving out of range of the signal. In the former it is optimal to delay the handover with the assumption that the signal will return while the latter benefits from an immediate handover as it is unlikely that the user will come back in range before the handover is complete.
By Mike Rahaim and Thomas Little
