Next generation PVB interlayer for improved HUD image clarity.
Head-up display (HUD) technology creates inherent driver safety advantages by displaying critical information directly in the driver's line of sight, reducing eyes off road and accommodation time. This is accomplished using a system of relay optics and windshield reflection to generate a virtual image that appears to hover over the hood near the bumper. The windshield is an integral optical component of the HUD system, but unfortunately the windshield-air interface causes a double image ghost effect as a result of refractive index change, reducing HUD image clarity. Current technology uses a constant angle wedged PVB windshield interlayer to eliminate double image at a single driver height. However, the HUD double image persists for all other viewing locations. Eastman Chemical Company has developed a new interlayer technology which eliminates the double image at all driver locations by tuning the wedge angle as a function of driver occupant seated height. The purpose of this paper is to discuss this exciting technology and its application to produce best in class HUD image clarity.
CITATION: Spangler, L., Hurlbut, J., Cashen, D., Robb, E. et al., "Next Generation PVB Interlayer for Improved HUD Image Clarity," SAE Int. J. Passeng. Cars - Mech. Syst. 9(1):2016.
Head-Up-Displays (HUD) technology uses a system of relay optics to create a transparent virtual image. This transparent virtual image provides the ability to display critical vehicle or safety information without requiring the user to look away from the road, hence the name Head-Up-Display. HUDs also have the inherent advantage that the user's eyes do not need to refocus from the real world scene to the HUD virtual image thereby reducing eyes off road time.  As shown in Figure 1 below, the HUD component includes various mirrors including at least one concave mirror. The concave mirror magnifies the HUD display creating a larger virtual image in front of the vehicle.
The windshield plays a critical role as a mirror in the HUD optical system to reflect the HUD image into the driver's Field-of-View (FOV). Fortunately for the driver but unfortunately for the HUD, the windshield is fundamentally not designed to be a mirror. The HUD windshield is not an ideal mirror because it is not fat and more importantly, is not very reflective.
Consider a practical example where an automotive windshield consists of two panes of soda-lime glass separated by a PVB interlayer. This construction produces two interfaces from air to glass and two reflections/refractions as shown in Figure 2 assuming that the PVB has the same optical material properties as the glass. Both interfaces produce reflections and refractions whose angles are governed by the Snell equation (1) shown below:
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)
Applying the Snell law to an air glass interface it is clear that the angle of refraction is more normal to the glass surface than the angle of incident. This phenomenon is especially important in terms of HUD optics because the refraction through the glass produces two paths from the driver's eye location to the same point on the display, meaning that the driver sees the same image in two different places as show in Figure 3. This second image is referred to as ghosting or double imaging and is shown in the grid pattern below as a blue grid compared to the red primary image grid.
Windshields used for HUD applications today use a technique called wedging to align both optical paths to the same point. This means that both optical paths from the display to the eye point start and end at the same 3D location superimposing the two images for a single design point as shown in Figure 4. This is accomplished by rotating the outer pane of glass through the use of a wedged PVB interlayer such that the primary image shown below in red and the secondary image shown below in blue overlap. For additional background refer to US Patent 5,013,134 .
This approach has a noticeable limitation in that the two different optical paths are only aligned at a single eye point location. Different drivers view the HUD image from different heights at different angles producing a double image for all driver heights except the height at which the wedge angle is optimized. This produces the ghosting anomaly as shown in Figure 5 and represents the limitation of current technology.
For the next generation of HUD technology the HUD image size, brightness and resolution will all increase. Naturally the goal is to improve HUD image clarity in line with the other HUD image improvements. However, the wedge angle cannot take multiple values at the same place on the windshield therefore any single point on the windshield cannot be optimized for viewing at all driver seated heights. The key to conquering this physical limitation is to consider that the intersection point to the windshield is different for different driver occupant seated heights as shown in Figure 6.
Using this approach, the wedge angle can be optimized for different linear distances from the nominal intersection point to ideally minimize ghosting at all driver locations as shown in Figure 7.
Calculating the angle required to eliminate ghosting for multiple seated height results in the "S-curve" wedge profile shown in Figure 8. This represents the ideal calculated wedge profile.
While the application of variable angle wedge to HUD windshields is theoretically expected to result in an exceptional improvement in HUD image quality at all eyebox locations, until now a myriad of practical challenges have prevented its successful implementation. Not only must the wedge angle be accurately controlled over small linear distances in the PVB interlayer which requires novel processing and good production control, but it must also be located precisely in the windshield during the lamination process. The following sections describe the creation and characterization of variable angle wedge samples. A sample specification, representative of one of the most difficult use cases in the marketplace, was chosen for this demonstration. In this paper, the samples are characterized and a measurement technique is presented to show the correlation between design and actual VAW image quality performance.
VARIABLE ANGLE WEDGE DESIGN
PVB interlayers designed for HUD applications typically have thickness profiles which create a constant wedge angle in the HUD zone. Typical wedge angles are between ~0.2mrads to 0.8 mrads depending on the make and model of the car. This value is traditionally a single constant value over the entire HUD region. Due to the issues discussed in the previous section, a need has been identified for materials incorporating defined changes to the wedge angle specifically in the HUD zone to reduce double image for all drivers. It is now possible to manufacture interlayers with a constant wedge or a variable angle wedge profile across the windshield height. This wedge angle can vary from high to low or low to high in response to wedge profile requirements. Examples of angle profiles that have been demonstrated are shown in Figure 9.
The ideal VAW "S curve" profile calculated for the target vehicle test case is shown in Figure 10. Ultimately, for this first investigation, the S curve was approximated by a linear rate of change profile. This change simplified VAW PVB production and assembly while minimizing deviation from ideal. In fact, the linear curve fit results in a maximum thickness deviation from the target profile in the HUD zone by only 0.0013mm or ~0.1%, (Figure 11).
In this study several windshields were produced with different PVB interlayer samples. The windshield thicknesses were measured across the height of the windshield using a Lumetrics Optigauge which uses the principle of interferometry to calculate layer thickness. Windshield thicknesses were measured approximately every 0.1 cm and wedge angles were determined by first smoothing the data using a Savitzky-Golay filter to increase the signal-to-noise ratio and then calculating the first derivative (i.e. delta y over delta x) in the HUD region. This allows data smoothing to reduce signal noise while maintaining angle details (Figure 12).
The wedge profiles of four sample windshields are shown below in Figure 13. This includes a windshield with a constant angle wedge profile (Sample P) and three windshields with variable angle wedge profiles (Samples B, H, F) having differing deviation from the target profile. It is possible to estimate the eyebox position along the linear length of the windshield for the short, nominal, tall and tall plus driver eyebox positions from the windshield design, installation, and location (Figure 6). These driver occupant heights are also indicated with respect to the wedge profile in Figure 13.
Sample P contains a constant angle wedge interlayer (~0.7mrad) which was optimized for the nominal driver. An example of the entire grid image monitored at the tall driver position is shown in Figure 14. For quick comparison, an excerpt from near the center of the grid image (red box) for each driver position is shown in Figure 15.
Based on the measured deviation of the angle profile of Sample P from the target optimized angle profile, one would expected that the HUD image quality would be good for the short, nominal, and tall driver, but poor for the tall plus driver. This agrees with the image quality observed in Figure 15.
Variable angle wedge windshield Sample B has the highest measured deviation from target profile in the nominal and tall eyebox regions but is closer to the target profile in the short and tall plus driver positions. Again, the measurement agrees with the observed image quality shown in Figure 16.
The angle deviation variation from target profile for VAW Sample H is larger for the short and nominal driver eyebox positions and gets closer to target for the tall and tall plus driver eyebox positions. Again, this measured deviation compares well to image quality shown in Figure 17.
Finally, the angle profile of the VAW Sample F is closest to the target profile for all locations and image quality in all locations is good as demonstrated in Figure 18. This level of image quality at all eye box locations is not achievable using a contast angle wedge profile.
Table 1 contains a qualtative assement of the visual quality for Sample P, B, H, and F and the deviation from the target profile. This table shows that there is good agreement between image quality and angle deviation, in most cases. Therefore, is it possible to achieve good image quality at all eyebox locations by varying the wedge angle through the HUD zone and creating a interlayer profile which is optimized.
As a result of the collaboration between Eastman, NSI and NSG it is clear that Variable Angle Wedge (VAW) PVB technology can be produced to even the most challenging target VAW profiles. Additionally, it is clear that variable angle wedge enabled windshields will improve HUD image quality especially for the extreme tall and short drivers. Based on
these findings, it is strongly recommended to further consider VAW technology for production intent evaluation.
[1.] Robb, E.R. and Cashen, D. (2014) Augmented Reality Human-Machine Interface: Defining Future Display Technology, 2014 Vehicle Displays and Interfaces Symposium p. 175-180
[2.] Smith, Ronald T. 1989. Ghost-Free Automotive Head-Up Display Employing a Wedged Windshield. U.S. Patent 5,013,134, filed September 28, 1989 and issued May 7, 1991.
Lora L. Spangler and Jeffrey Hurlbut
Daniel Cashen and Emily Robb
NS International Ltd.
Eastman Chemical Company
730 Worcester Street
Springfield, MA 01007
HUD - Heads Up Display
VAW - Variable Angle Wedge
Table 1. Qualitative evaluation of Image Quality versus Angle Delta from Target Angle. Average image quality is determined by average separation distance between primary image and reflected image from top to bottom of grid slice. Short Nominal Angle Delta Angle Delta Avg. Image from Target Avg. Image from Target Sample ID Quality (>10 to 15%) Quality (>10 to 15%) P OK OK B OK OK X H NOK X NOK X F OK OK Tall Tall Plus Angle Delta Angle Delta Avg. Image from Target Avg. Image from Target Sample ID Quality (>10 to 15%) Quality (>10 to 15%) P OK NOK X B NOK X OK H OK OK F OK OK
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||polyvinyl butyral for head-up displays|
|Author:||Spangler, Lora L.; Hurlbut, Jeffrey; Cashen, Daniel; Robb, Emily; Eckhart, Jim|
|Publication:||SAE International Journal of Passenger Cars - Mechanical Systems|
|Date:||Apr 1, 2016|
|Previous Article:||Reliability evaluation of thin, lightweight laminates for windshield applications.|
|Next Article:||Vehicle longitudinal control algorithm based on iterative learning control.|