High-Resolution LCD Headlamps for Intelligent Lighting
Vehicle headlamps that adapt to external conditions are established and available in mass-production vehicles; now the market is requesting headlamp systems with increasing resolution. Active-matrix LCD (AMLCD) with amorphous silicon thin-film transistors (a-Si TFTs) is a mature and cost-efficient technology that is known for its reproducible and homogeneous performance. Integrating an AMLCD with a-Si TFTs into an LED-based headlamp can fulfill all specifications for realizing a fully adaptive headlamp setup with high resolution.
by Christiane Reinert-Weiss and David Duhme
SINCE the first candlelit headlamps were mounted on horse-drawn carriages, lighting technology has evolved alongside the vehicles in which it is installed. From acetylene gas lamps in the late 1880s to xenon headlamps in the 1990s, lighting modules improved in efficiency but remained analog. The implementation of adaptive systems in modern vehicles creates a new performance demand
on headlamps that traditional analog technologies cannot fulfill. Whereas traditional analog headlamps simply illuminate their environment with a predefined light distribution, the new digital systems react to the environment and require that the headlamps dynamically adjust their beam patterns appropriately. To be able to do this, the headlamp system uses data like speed, steering angle, GPS, and ambient data provided by sensors and cameras already integrated in modern vehicles. Based on this data set, methods like object recognition calculate an optimal adapted-lighting distribution, which is then projected onto the street in real time. Apart from dynamic bend lighting, which improves illumination during cornering, the glare-free high beam is the most prominent feature implemented in state-of-the-art headlamps (see Fig. 1).
Fig. 1: The above illustration shows a car with adaptive headlamps employing a glare-free high beam. Source: HELLA GmbH & Co. KGaA.
The era of intelligent headlights in series-production vehicles began with the “High Beam Assistant” introduced by BMW in 2005, which automatically turned off the xenon-based high beam when oncoming traffic was detected. Mercedes followed with its “Adaptive High Beam Assist” in 2009, which adapted the range of the low beam automatically, and VW introduced its “Glare Free High Beam” in 2010, using rotating drums to create vertical cut-off lines. The latter two systems both use a xenon light source. The first matrix LED headlamp acting as a solid-state technology for glare-free high beams was introduced by Audi in 2013.
From an automated “on” and “off” to constantly adapting beam patterns, each consecutive generation of adaptive headlamp modules increased the number of realizable functionalities to improve both safety and convenience for the driver. Today’s headlamps can create arbitrary light distributions like de-glaring other traffic or reducing the illumination of reflective traffic signs to avoid self-glaring while keeping the surroundings fully illuminated. To accomplish this, the headlamp modules divide
the illuminated area in front of the vehicle into a grid of small segments with individually controlled light intensity. The higher the resolution of this grid, the better the realizable lighting functions. A state-of-the-art adaptive LED-matrix technology is represented by the “Multi Beam LED” implemented by Mercedes Benz in 2016, which divides the area in front of the vehicle into 84 light segments.
Digital Light: Current Technologies
The above description of an adaptive headlamp very much resembles the description of an active-matrix liquid-crystal display (AMLCD). The active matrix is a grid of pixels, and the amount of light passing through each pixel is actively controlled by drivers. But can an AMLCD headlamp compete with state-of-the-art technologies? At the moment, the three other main technologies3 that are being investigated and implemented in the search for higher resolution and optimized functionality are:
µAFS (Micro Adaptive Frontlight System): By increasing the number of LEDs on a single chip to 1,024, headlights with ≥1 K individually controllable pixels each can be realized.
DMD (Digital Mirror Device): A strong LED light source illuminates a chip with >500 K micro-mirrors.
Laser Scanner: A laser beam is directed via a mirror onto a phosphorous plate to create arbitrary lighting distributions.
Regardless of the method used to create a high resolution, all of the systems mentioned need additional optics to create the desired light distributions. Compared to these technologies, an AMLCD headlamp can not only similarly fulfill the requirements of a headlamp application, but provide distinctive advantages to be considered for further investigation.
Advantages of an AMLCD Headlamp with LED Matrix
In a headlamp using matrix LEDs alone, like the “Multi Beam LED” from Mercedes Benz mentioned earlier, or the µAFS, the number of pixels equals the number of LEDs implemented in the matrix. Therefore, the number of achievable pixels is limited by the number of LEDs one can operate within the limited installation space of a headlamp or integrate on one chip. Both approaches therefore result in a limited resolution. By combining an LED-matrix backlight with an AMLCD module, the number of individually addressable pixels can be increased to tens of thousands while at the same time decreasing the number of required LEDssubstantially. The LED matrix for the AMLCD headlamp developed by the authors’ team,1 as described below, consists of only 25 LEDs. The electronic control unit (ECU) can drive the LEDs at a maximum of 3 amps. For a high-beam light distribution, the LCD module consumes 75 watts.
In contrast to DMDs or laser scanning, an AMLCD headlamp does not need movable mirrors or other mechanical elements. AMLCDs are already mass produced for flat screens and are available at low cost. For these reasons, the Institute for Large Area Microelectronics (IGM or Institut für Großflächige Mikroelektronik) at the University of Stuttgart and HELLA GmbH & Co. KGaA investigated the feasibility of an LED-based AMLCD headlamp as part of the BMBF (Bundesministerium für Bildung und Forschung, German Federal Ministry of Education and Research) project VoLiFa2020.2
Development of an LED-Based AMLCD Headlamp
When considered against the many possible combinations of light sources and light-distribution mechanisms, LED-array background lighting with an AMLCD module offers many advantages. LED matrix is a technology already used as a light source in headlamps, as mentioned above. In addition to LED technology’s well-known benefits, such as compactness and energy efficiency, its highly adaptable light spectrum makes it an ideal light source for AMLCD modules. Halogen and xenon lamps both have light spectra with a UV portion. As the organic components of an LC display, particularly the liquid-crystal composition, degrade under UV illumination, choosing a well-adapted light source increases the lifetime of the system significantly. White high-power LEDs such as phosphor-coated GaN LEDs are able to provide a sufficient luminous intensity and the necessary UV-free light spectrum for a durable AMLCD headlamp.
But integrating an AMLCD-module into a headlamp raises problems, the most substantial of which are the optical efficiency and stability of the module under thermal and optical stress. For an automotive application, the module must work reliably in the temperature range of –40°C up to +125°C. At the same time, it must be able to withstand illuminances of ≥20 Mlx provided by the LED-matrix background lighting for several hours per use for the life span of a vehicle, which is considered to be ~15 years or 300,000 kilometers. To adhere to the regulations, a high beam should provide an illuminance of >120 lx on a surface 25 meters in front of the vehicle when turned on. When turned off, oncoming traffic at the same distance should not be exposed to more than 1 lx (based on both headlamps) to avoid glare. To fulfill these boundary conditions, a headlamp needs a contrast ratio >240:1. In order to ensure this ratio, the contrast of an AMLCD integrated in a headlamp should be considerably higher than that.
For an optimal contrast ratio, a twisted-nematic (TN) cell with perpendicular polarizers is a good possibility. Active-matrix TN-LCDs are a robust technology and easily adapted for headlamp applications. The limited viewing angle of a TN-cell is uncritical in this context, as the AMLCD is used for projection instead of displaying information. However, the thermal stability of standard liquid-crystal compositions used in flat screens does not encompass a nematic phase spanning at least 165°C between crystallization temperature and clearing point. Therefore, Merck KGaA developed a novel liquid-crystal composition specifically for this application.1
The choice of a suitable thin-film transistor technology for the AMLCD backplane is driven by the need for a reliable, cost efficient, and easily available technology. Amorphous silicon transistors (a-Si TFTs) are the most mature technology for AMLCDs and available at low cost. Considering their sensitivity to illumination, a-Si TFTs did not at first seem
to be suitable for an application in headlamps. Typical input characteristics of a-Si TFTs under different lighting conditions are shown in Fig. 2(a). When the channel area is exposed directly to >20 Mlx illuminance, the input characteristics are severely impaired compared to measurements done under ambient light. However, as shown in,1 a metallic light shield was applied, which improved the input characteristics sufficiently, even under 40 Mlx illuminance, to make a-Si TFTs applicable in a headlamp module. Such a TN AMLCD with a-Si TFT backplane can be seen in Fig. 2(b).
Fig. 2: At left (a) appear typical input characteristics of a-Si TFTs under different lighting conditions and at right (b), a bonded TN AMLCD with an a-Si TFT backplane. Source: IGM, University of Stuttgart.
In a typical AMLCD embodiment, at least 50 percent of the background lighting is lost in the first polarizer. This loss has to be decreased substantially to make an AMLCD-LED headlamp system that can compete with existing solutions. But in this case, the source of the problem is also the key to its solution. Standard display polarizer foils do not provide
sufficient robustness against light and temperature stress to be usable in this application. Therefore, the AMLCD module uses metallic wire grid polarizers instead. The solution to decreasing the losses implemented for the VoLiFa2020 headlamp1 takes advantage of the properties of those wire grid polarizers to separate the background lighting into its two planes of polarization. Each polarization is then directed through its own active-matrix area. This results in two discrete optical paths, one for each plane of polarization (see Fig. 3). By this means, the optical yield of the generated light after the analyzer increases up to 80 percent. The two optical paths are then recombined by secondary optics to create a homogeneous projection on the street 25 meters in front of the vehicle. This approach has the additional advantage of redundancy, allowing the system to compensate for pixel failures in one AM area with the information provided by the other.
Fig. 3: The AMLCD high-beam design incorporates two discrete optical paths, one for each plane of polarization. Source: HELLA GmbH & Co. KGaA.
The thermal management of the 25 LEDs, ECU, and optical components like the LCD and polarizers is done by an active cooling system. The system design is based on computational fluid dynamics simulations and includes a radial fan and air-guiding elements. The major obstacle was to balance the system regarding package space, thermal management, optical performance, power consumption, efficiency, styling, resolution, functionality, and field of view (Fig. 4).
Fig. 4: Field of view – Each headlamp illuminates an area of 30° width and 10° height in front of the vehicle, overlapping in the central area. The 30K switchable pixels per headlight result in a resolution of 0.1° × 0.1°. Source: HELLA GmbH & Co. KGaA.
By using polarizers, separating the planes of polarization, and choosing a suitable backplane technology and a dedicated twisted-nematic (TN) liquid-crystal composition, an AMLCD high beam with LED backlighting can be made readily capable of providing ≥30 K switchable pixels per headlight to create fully adaptive lighting. The contrast ratio of the system
implemented in the VoLiFa2020 headlamp was measured at up to 490:1, surpassing the minimum requirement of 240:1 considerably. The results were presented at Display Week 2017 in Los Angeles, earning a student paper award and the I-Zone Best Prototype Award. The VoLiFa2020 headlamp has been integrated in the headlamps of a Porsche Panamera test vehicle (Fig. 5), and perception studies are currently being conducted to evaluate and adjust new functionalities as a part of the project VoLiFa2020.1
Fig. 5: At left (a), the AMLCD high beam with LED-matrix background lighting has been incorporated into a Porsche Panamera headlamp. At right (b), that headlamp’s projection at a 10-meter distance is shown. Source: HELLA GmbH & Co. KGaA.
At the moment, the number of pixels that could easily be realized with an AMLCD surpasses the number of pixels that can be fed with real-time data computed from the input of the vehicle’s sensors. This certainly will change in the future. However, the benefit of an increasing number of pixels and higher resolution reaches its limit at the threshold of human perception. The size and distortion of the projection of a single pixel as well as the unevenness of the projection area influence this perception. Further studies will show at which point the limit of added value per added pixel is reached.
The new generation of headlamp systems creates non-analog light to provide fully adaptable lighting to increase the safety and comfort of the driver. Lighting functions like de-glaring other traffic and road signs, projecting distance warnings and navigation onto the street, and indicating safety zones for other road users as well as the width needed to pass other traffic safely (see Fig. 6) – to name some of the more prominent ones – can be realized simultaneously and in real time. Adding even more flexibility, specific light distributions for different requirements can be implemented without changing ray optics. Lighting functions can be simply added, modified, or disabled by a simple software update to adapt to the lighting scheme of specific car brands, consumer preferences, regional conditions, or changing legislation.
Fig. 6: Possible functionalities of adaptive headlights include indicating passing lanes (a) and bicycle safety zones (b). Source: HELLA GmbH & Co. KGaA.
Considering this, it is no wonder that the market share of fully adaptive headlamps has been rising steadily. Different approaches to creating a high-resolution headlamp that can adapt illumination with the least possible delay to ever-changing situations on the road are currently being investigated. An AMLCD module with LED-matrix background lighting is a state-of-the-art solution aimed at meeting these challenges.
1C. J. Reinert-Weiss, H. Baur, S. A. A. Nusayer, D. Duhme, and N. Frühauf, “Development of Active Matrix LCD for Use in High-Resolution Adaptive Headlights,” JSID 25, 90–97, 2017 (onlinelibrary.wiley.com/doi/10.1002/jsid.534/full).
2BMBF: Project “Volladaptive Lichtverteilung für eine intelligente, effiziente und sichere Fahrzeugbeleuchtung (Fully adaptive
light distribution for an intelligent, efficient, and safe car lighting, VoLiFa2020),” https://www.photonikforschung.de/foerderung/beleuchtung-und-led/projekt/volifa2020.html.
Dipl.-Ing. Christiane Reinert-Weiss received her diploma in electrical engineering and information technology from the University of Stuttgart in 2008. She is currently a Ph.D. student at the Institute for Large Area Microelectronics at the University of Stuttgart. She can be reached at firstname.lastname@example.org. From 2008 to 2012, David Duhme studied mechanical engineering at the University of Applied Sciences Südwestfalen, in cooperation with the automotive supplier HELLA GmbH & Co. KGaA. Since 2012 he has been developing high-definition headlamps in the predevelopment department of HELLA as a research engineer.