This article was published in the April 2013 issue of LEDs Magazine.
View the Table of Contents and download the PDF file of the complete April 2013 issue, or view the E-zine version in your browser.
Of the many design challenges facing LED-based solid-state lighting (SSL) applications, perhaps there is none greater than that of expectations. There are expectations around the application. There are expectations around the incumbent technology. There are expectations around the way it has always been done, and, as a result, there are expectations around the way it should be done going forward. What if we were able, however, to design with a clean sheet of paper? Take roadway lighting as an example. If we were to take that application, deconstruct it, and come at it from a different angle, what might we do differently, and how are LEDs specifically suitable tools in this redesign?
When we think about the job of lighting a roadway, we are conditioned to think about what is happening right in front of us. We think about targets in the road and response time in identification. In fact, the entire series of metrics for roadway lighting is modeled around these requirements. From this standpoint, our examination of roadway lighting is fundamentally no different than our examination of office lighting. The conditions and demands of the tasks, however, couldn’t be more different.
Once we step back, one of the things we can appreciate regarding roadway lighting is that we are invariably talking about night-time situations. While the human visual system has an amazing ability to tolerate a wide range of conditions, the mechanisms that allow for those ranges vary for different lighting levels — night-time environments especially. To better appreciate how those mechanisms come into play, we need to consider the retina and its component parts.
The retina is incredibly complex, but its basic role can be summarized by two types of photoreceptors: cones and rods. Cones are located predominantly in the center of the retina in the fovea. Rods, which greatly outnumber cones, surround the fovea and encompass the periphery of the retina. The retina is in simplest terms a camera. It produces images for the central nervous system (CNS) to interpret.
The CNS-to-photoreceptor pathways best define the photoreceptor’s role in vision. Each cone, in effect, has its own direct path to the CNS. A quanta of information is personally escorted to the brain for processing. This one-to-one relationship defines its role in higher order perception such as fine detail discrimination and color analysis. The peripheral vision pathways to the CNS are shared by large groups of neighboring rods. Light that grazes one edge of the group triggers a response on the far edge. Through this mechanism, rods preform their basic role of gross peripheral motion detection.
Using night-time driving as an example of the mechanism, our eyes are directed for the majority of time at the roadway, where the cones are aiding in the analysis of detail. When something appears in the periphery, say a deer approaching the shoulder of the road, this sight registers across many groups of rods, signaling movement to the CNS. At this point, the eyes move and perhaps the head pivots, so that the cones can be engaged for better detail analysis and subsequent reaction.
Rule #2: Appreciate the importance of peripheral detection in night-time driving.
Our current metrics are concerned with foveal vision exclusively, yet the fovea takes up a tiny percentage of the visual field. We essentially light the road as depicted in Fig. 1. Mark Rea, director of the Lighting Research Center and professor at Rensselaer Polytechnic Institute, has written extensively on the subject. Rea has said that considering just the fovea in driving is akin to driving while looking down a long, narrow tube. Given the choice, would we choose the field of vision on the inside of the tube or the outside in order to drive? While what is inside the tube is important, this example illustrates that the outside of the tube — our peripheral vision, at the very least, deserves some consideration.
While rods work in groups, they are individually much more sensitive to light than cones. Able to absorb and register even a single photon, one immediately sees their advantage in night-time conditions. Indeed, as light levels drop, the rod-to-cone activation ratio increases until rod sensitivities are at a peak level in night-time conditions.
Where the spectrum of light is concerned, the rods and cones respond similarly to higher wavelengths. Rods are, however, much more sensitive than cones to lower wavelengths, especially after they have time to adapt to night-time conditions. If one of our goals is to optimize the lighting to better aid in peripheral target detection, we should be working with a spectrum that is optimized to that task and optimized to the photoreceptors (rods) engaged in that task.
Rule #4: Eliminate double work.
Regardless of the importance of peripheral vision, we still need cones for sign identification/reading and analysis of detail in the roadway. The metric that matters, just as in office lighting, for example, is contrast. How do we present the task in proper relief? Strong forward lighting (such as provided by car head lamps) with narrow optics will optimally illuminate the vertical plane and present a snappy, sharp shadow with an excellent dichotomy between light and dark. Current roadway metrics, mostly concerned (again, like office lighting) with horizontal illumination, don’t even consider the vertical plane. As written, the application requirements only consider overhead lighting, which can have a deleterious effect on contrast when combined with forward lighting on cars. Roadway lighting needs to complement forward lighting on automobiles and aid in the creation of contrast and clear, decipherable indicators to which our CNS can respond.
Rule #5: Light the edges.
More importantly, however, is the ability to identify hazards prior to them being in the roadway. Rea has suggested, only partially in jest, that better viewing conditions may be gained by simply pivoting roadway lighting 180⁰ in order to light the shoulder (Fig. 2). The job of lighting the roadway is then left to headlights. The optimal solution is most likely a combination of that approach and current practices, but the clues are there.
The issue with incumbent technology in roadway applications is the one-size-fits-all limitations. We start with a high flux, high wattage, omnidirectional light source, and we attempt to corral the beam to meet the application. The approach is inherently inefficient from an optical perspective. There is no opportunity for nuance or spectral shaping.
With LED point sources, we build a fixture piece-wise until we have the perfect distribution — no more; no less. As Fig. 3 shows, SSL fixtures can be designed to produce almost no light behind the poles. Through proper binning, we are able to spectrally shape the output in order to best match the visual needs. In the example we have been using for roadway lighting, we can imagine many different designs or a combination of attributes in one package.
We could have a component of the beam that lights the shoulder and surrounding areas of the roadway for the optimal spectrum of the rods. We could concurrently light the roadway with another spectrum ideal for foveal vision and contrast. We could have peripheral lighting that stays on constantly in rural settings or in areas of high deer traffic. Conversely, thanks to SSL instant start capabilities, we could have peripheral lighting that comes on as a function of peripheral motion.
The options are open-ended. What is clear is that new technology allows designers the opportunity to not only work with new tools but also return to the applications themselves and rethink the way things are done. When we do that, the value of lighting is optimized in its abilities to help people. We escape the morass of expectations, and we evolve as an industry.