Head-up display, method and optical element

Abstract

A stereoscopic Head-Up Display, HUD, is disclosed. The stereoscopic HUD comprises image sources for generating images, a pupil-tracking system configured to determine left and right pupil positions of a viewer, and an optical element positioned between the plurality of image sources and a viewing position. The optical element is configured to split an incident ray into multiple output rays at respective different angles, and configured such that output rays split from a single input ray are not incident on both left and right pupil positions at the viewing position. The stereoscopic HUD also comprises a controller configured to cause a first image source to display an image point corresponding to a first image for the left pupil and to cause a second image source to display at least an image point corresponding to a second image for the right pupil.

Description

[0001] HEAD-UP DISPLAY, METHOD AND OPTICAL ELEMENT

[0002] Technical Field

[0003] The present invention relates to a head-up display (HUD), a method for a HUD and an optical element. In particular, the present invention relates to a HUD that can provide large field of view (FoV) imagery to a viewer.

[0004] Background

[0005] Head-up displays (HUDs) are display systems that project images directly into a viewer's line of sight, allowing them to view the images without looking away from their usual viewpoints. HUDs superimpose an image on top of the environment and have various applications, including automotive, aviation, and augmented reality (AR).

[0006] It is desirable to have a very large FoV for a HUD. For example, in automotive settings, filling a full windshield with HUD content can allow new applications, such as directions appearing to be superimposed on the road in front of a viewer. This requires a display having around a 40x20 degree FoV, which is challenging to achieve because of the extremely large etendue at the viewer. For example, for a typical 160x160 mm eyebox, combined with a 40x20 degree FoV requires around 6000 mm2of etendue, compared to a typical micro-display which may only handle a few hundred mm2of etendue. Increasing the etendue of such a display requires a large amount of space, too large to be easily integrated into the space available.

[0007] It would therefore be desirable to provide a HUD having a large FoV where the HUD system is more compact.

[0008] Summary

[0009] The present disclosure addresses the above desire. According to a first aspect, there is provided a stereoscopic Head-Up Display, HUD. The HUD comprises a plurality of image sources for generating images, a pupil-tracking system configured to determine left and right pupil positions of a viewer, and an optical element positioned in an optical path between the plurality of image sources and a viewing position. The optical element is configured to split an incident ray into multiple output rays at respective different angles. The optical element is further configured such that output rays split from a single input ray are not incident on both of the left and right pupil positions at the viewing position. The HUD further comprises a controller configured to cause at least one first image source of the plurality of image sources to display at least an image point corresponding to a first image for the left pupil and to cause at least one second image source of the plurality of image sources to display at least an image point corresponding to a second image for the right pupil, the optical element and the configuration of the plurality of image sources combining such that rays of the first image are received at the left pupil position and rays of the second image are received at the right pupil position.

[0010] The optical element may form part of a see-through combiner. The action of the optical element is similar to that of a diffuser. Suitable optical elements having the properties above include holographic optical elements (HOEs), polarisation volume gratings, surface relief diffractive optical elements. The defined properties of the optical element allow for replication of an incident ray without translation of the incident ray along a length of the optical element. Translation of rays may be undesirable due to the optical element essentially acting as a waveguide, manufacture of such a waveguide having a suitably large size, possibly also with curvature, may be challenging. In contrast, the optical element discussed herein may be easier to manufacture and use.

[0011] An optical element configured to split, or replicate, single input rays into respective multiple output rays at different angles allows for an increase in etendue of the display by generating a plurality of replications of the input light. In this way, the HUD can provide a large FoV without the need to provide an infeasibly large optical engine form factor because the output rays corresponding to a single input ray can cover a wide area at the viewing position.

[0012] The pupil-tracking system and controller allow the HUD to control the rays that are incident on each pupil. In other words, the HUD is a display system that prevents binocular crosstalk between different images. This may be particularly useful where the first and second images represent stereoscopic or binocular images of a single scene such that the viewer perceives a three-dimensional scene via stereoscopy. Since rays from a single image are not incident on both pupils at any one time, the HUD is capable of ensuring that the rays that are incident on each pupil represent the correct image for that pupil.

[0013] Each image source may be configured to generate at least part of an image at a given time. More particularly, each image source may be configured to generate one or more image points that form the image(s) received by the viewer. In other words, and as will become clear from the following discussion, each image source may be configured to contribute to one or more respective images. In some examples, each image source may be a separate optical module (sometimes referred to as a picture generating unit). In other examples, at least one of the optical modules may be configured to generate two or more image sources by timemultiplexing the image sources at a sufficient speed such that their outputs are perceived by the viewer as a single image due to persistence of vision. Each image source may be configured to generate light at one wavelength, or a fixed plurality of wavelengths (corresponding to red, green and blue light, for example).

[0014] Where an image source is configured to generate light at a fixed plurality of wavelengths, the light at each wavelength may be time-multiplexed so that a viewer perceives a colour image. Furthermore, where light at a fixed plurality of wavelengths is used, there may be an optical element corresponding to light of each wavelength due to the wavelength dependence of the action of an optical element, such as a HOE.

[0015] The pupil-tracking system and controller allow the HUD to update the content generated by the image sources to ensure that the viewer receives the appropriate content at each pupil. This may also allow the HUD to simulate parallax as the viewer’s head moves by recalculating and displaying a view that matches the viewer’s current eye position.

[0016] The optical element may be configured such that the multiple output rays split from an incident input ray form multiple discrete bands of output rays (due to the ray-splitting property). The union of bands for a single input ray may be referred to as a light footprint. The optical element and the image sources may be configured such that bands corresponding to rays from each of the image sources are substantially space-filling at the viewing position. This may help to ensure that images are always received at both pupils, further enhancing the HUDs ability to generate a large FoV In some examples, this may involve the bands being substantially tessellated at the viewing position. In some other examples, neighbouring bands (corresponding to input rays from different image sources) may at least partially overlap.. This space-filling criterion may be imposed based on the geometry of the viewer’s pupils, and may require three or more image sources. In some examples, a minimum distance between adjacent bands corresponding to a single input ray may be greater than 75mm, for example 78mm, or 85mm. In examples, a width of a band at the viewing position may be less than 50mm, such as 48mm or as small as 38mm. These dimensions may be adjusted based on a particular viewer, or set of viewers. In examples where neighbouring bands at least partially overlap, these dimensions may be adjusted. For example, the width of a band at the viewing position may be greater than 50mm and / / or the minimum distance between adjacent bands corresponding to a single input ray may be smaller than 75mm. In some examples, the optical element may be configured such that multiple output rays split from a single input ray are not incident on a single pupil position at the viewing position. This is a tighter constraint on the optical arrangement of the HUD. If two or more output rays replicated from a single input ray are received at a single pupil, this will give the impression that the imaged is focussed at a position of the optical element since those two or more output rays have diverged from a common point on the optical element. In other words, the generated image has a focal depth at the optical element. This additional constraint on the optical element therefore allows the generated images to have a focal depth away from the optical element.

[0017] Having the focal depth of the images beyond the optical element may reduce certain known phenomena related to augmented reality (AR) systems such as focal rivalry, vergenceaccommodation conflict (VAC), and parallax error with head motion. For example, setting the focal depth of the images at a distance approximately equivalent to one or more objects in the environment beyond the optical element may reduce the effects associated with a mismatch between the focus of those one or more objects and the images generated by the HUD.

[0018] As mentioned above, the optical element may form multiple bands of output rays for each input ray. In some examples the optical element and the image sources may be configured such that when an image source illuminates a region of the optical element, a plurality of discrete sub-bands of output rays are formed within a band.

[0019] The optical element may be configured such that a minimum distance between nearest subbands corresponding to a single image source, at the viewing position, is greater than 6mm. This roughly corresponds to the width of a dilated pupil, but may be dependent on the particular viewer and so may be adjusted accordingly.

[0020] This stronger constraint on the images means that the images are naturally smaller in area at the viewing position compared to the images possible without this constraint. This may mean that more images are required to fulfil a space-filling criterion at the viewing position. For typical pupil widths and interpupillary distances, there may be nine or more sets of subbands. The stronger criterion can then be expressed as the gap between sub-bands of the same type being greater than the size of a viewer’s pupil. The union of the sub-bands corresponding to each of the image sources substantially tessellates at the viewing position. For instance, where each sub-band forms a circle at the viewing position, it may be advantageous to use a hexagonal tessellation of the sub-bands due to the optimal circle packing property. Other options are possible such as a square or triangle tessellations, using circular or other shapes. In some examples, a sub-band corresponding to the first image source may at least partially overlap a sub-band corresponding to the second image source. As discussed, the optical element may comprise one or more HOEs. HOEs may be advantageous for automotive HUDs since they have good reflective efficiency for a design wavelength and incident angle, whilst having little interaction with transmitted light. This may allow the windscreen, and the HOE which may be mounted onto the windscreen, to remain transparent for nearly all wavelengths and incident angles, thereby improving safety.

[0021] According to a second aspect of the present invention, there is provided an automotive HUD comprising: a windscreen, and the HUD according to the first aspect, wherein the optical element of the HUD is coupled to the windscreen. Where the optical element comprises an HOE, this may be mounted to the windscreen. In an automotive setting, there may be additional windows and reflective elements present that do not form part of the HUD. In such cases, it may be possible to configure the optical element such that unintended reflections of the image channels are not directed to the viewing position. This may be possible by blocking certain undesirable ray paths and / or ensuring that rays exterior from the automotive system are not coupled in in a direction substantially towards the viewing position.

[0022] According to a third aspect of the present invention, there is provided a method comprising: determining a position of a left pupil of a viewer; determining a position of a right pupil of the viewer; determining a first image for viewing by the left pupil; determining a second image for viewing by the right pupil; determining at least one first image source for displaying at least part of the first image based on the position of the left pupil; determining at least one second image source for displaying at least part of the second image based on the position of the right pupil; and displaying at least an image point corresponding to the first image by the at least one first image source and the displaying at least an image point corresponding to the second image by the at least one second image source via an optical element positioned in an optical path between the first and second image sources and the positions of the left and right pupils, wherein the optical element is: configured to split an incident ray into multiple output rays at respective different angles, and configured such that output rays split from a single input ray are not incident on the positions of the first and second pupils simultaneously. The method according to the third aspect may share any of the above features with the HUD according to the first aspect or the automotive HUD according to the second aspect.

[0023] According to a fourth aspect there is provided an optical element for a stereoscopic HUD, the optical element comprising a surface feature configured to replicate an incident ray into a plurality of discrete output rays at different angles, wherein: the angles of the discrete output rays are dependent on a property of the incident ray, and the angles of the sparse replications are such that the discrete output rays are not incident on the positions of first and second pupils at a viewing position simultaneously. When forming part of a stereoscopic HUD, the surface feature may be an exterior surface feature, or it may be embedded or sandwiched. For example, the surface feature may be on, or coupled to, one or more interior surfaces.

[0024] The third and fourth aspects may also include any of the features discussed with reference to the first aspect.

[0025] Further features and advantages of the invention will become apparent from the following description of preferred embodiments of the invention, given by way of example only, which is made with reference to the accompanying drawings.

[0026] Brief Description of the Drawings

[0027] Figure 1 shows a HUD system according to an example;

[0028] Figure 2 shows a top-down view of a display arrangement for a single input ray according to an example;

[0029] Figure 3A shows a top-down view of a display arrangement for three input rays according to an example;

[0030] Figure 3B shows a front view of the example shown in Figure 3A;

[0031] Figure 4 shows an optical architecture according to an example;

[0032] Figure 5 shows a top-down view of a display arrangement for a single input ray according to a further example;

[0033] Figure 6 shows the action of an optical element on a single image point from an image source according to an example;

[0034] Figure 7 shows areas of an eyebox corresponding to nine independently controllable input channels according to an example;

[0035] Figure 8 shows an illustration of a viewer’s pupil locations relative to the example shown in Figure 7;

[0036] Figure 9 shows an optical architecture according to a further example;

[0037] Figure 10 shows a method according to an example;

[0038] Figure 11 shows a front view of a display arrangement for three input rays according to a further example; and

[0039] Figure 12 shows areas of an eyebox corresponding to independently controllable input channels according to a further example. Detailed Description

[0040] As has been briefly described above, a large FoV HUD has many potential applications. Conventionally, for a HUD to provide a large FoV, the HUD would require a correspondingly large form factor. This is impractical in most applications. The inventors have found that optical elements having certain properties may be used to increase the etendue while keeping the size of the HUD system small. A HUD according to this disclosure makes use of an optical element that, within a geometric optics representation, splits a single input ray into multiple output rays at different angles that can be directed towards an eyebox. In this way, splitting a single input ray into n output rays results in an / / -fold increase in the etendue of the display. As used herein, an “eyebox” defines pupil positions in which an image can be viewed; the volume in which a viewer’s pupil can be positioned to view an image, which may be a composite image formed from one or more image sources.

[0041] Fig. 1 shows, in general terms, a HUD system 100 that utilises this concept. The HUD system 100 may be configured to generate stereoscopic images to a viewer with a large FoV. A large FoV can be achieved from an optical arrangement of the HUD system 100 that ensures that images may always be received by a viewer, regardless of viewer head position. The HUD system 100 comprises a plurality of image sources 102. The image sources 102 are configured to generate light at one wavelength, or a plurality of wavelengths (corresponding to red, green and blue light, for example). The image sources may, for example, comprise a laser module or an LED and a possibly also modulator to produce an image by modulating the light, such as an LCD or DMD.

[0042] The HUD system 100 further comprises a combiner 104, such as a see-through combiner. The combiner 104 comprises an optical element 106, as will be discussed in further detail below, and a transparent substrate 108. The combiner 104 receives HUD content while providing a view of an environment beyond the combiner 104. The transparent substrate 108 may be adapted to the specific implementation. For example, where the HUD system 100 is part of an automotive HUD, the transparent substrate 108 may comprise a windscreen, also referred to as a windshield. Where the HUD system 100 is part of an aviation HUD, the transparent substrate 108 may comprise a standalone screen configured to be positioned between a pilot and a canopy or be provided by the canopy itself. In both cases, the transparent substrate 108 may be made from a suitable transparent material such as laminated glass, tempered glass, polycarbonate, acrylic, or advanced composites. “Large FoV” as used herein may be dependent on the particular application. For example, the achievable FoV of the HUD system 100 may be limited only by the size of the combiner 104. Where the combiner 104 comprises a windscreen, the HUD system 100 may be capable of filling the full windscreen with image content (e g. a 40x20 degree FoV).

[0043] Such large FoVs are achievable because the HUD system 100 further comprises an optical element 106 that is coupled to the transparent substrate 108. The optical element 106 is configured to split a single input ray into multiple output rays, as will be discussed in further detail below. The coupling of the transparent substrate 108 and the optical element 106 may involve the optical element 106 being mounted onto the transparent substrate 108, or the optical element 106 comprising certain properties and / or surface features of the transparent substrate 108. The type of coupling may be determined by the type of optical element 106 used. In some examples, the transparent substrate 108 and the optical element may be formed as a single component.

[0044] The HUD system 100 is arranged such that images generated by the image sources 102 are directed by the combiner 104 towards a viewer’s pupil 108, such that the viewer can see the images while still being able to view the environment beyond the combiner 104.

[0045] The HUD system 100 further comprises an pupil-tracking system 112 configured to determine a location of one or more of a viewer’s pupils 110. The pupil-tracking system 112, also referred to as an eye-tracking system, may be coupled to a controller 114 and provide the controller 114 with data indicating the location of the viewer’s pupil(s) 110. The pupil-tracking system 112 may provide periodic data, such as a data at regular time intervals or aperiodic data, such as data provided when something changes, indicating the location of the viewer’s pupil(s) 110 as a function of time. The pupil-tracking system 112 may be used in conjunction with control of the image sources 102 to update the content of the images received at the viewer’s pupil(s) as the viewer’s pupil(s) and / or head moves. In particular, appropriate content may be provided to each of the viewer’s pupils 110 such that the viewer perceives a single three- dimensional view via stereoscopy. The controller 114 may be configured to control the image sources 102 based on data received from the pupil-tracking system 112. For example, the controller may control appropriate image sources to direct a left image to a left pupil and a right image to a right pupil through the action of the optical element.

[0046] Stereoscopic imagery is a technique used to create an illusion of depth in an image by presenting two images from different viewpoints separately to left and right pupils 110 of the viewer. For stereoscopic imagery, output rays intended for a left pupil should not be incident on a viewer’s right pupil and vice versa. To achieve this, the optical element 106 is designed or configured to scatter light into particular region(s) or area(s), as will be discussed herein. The present disclosure uses an optical element 106 that splits a single input ray into multiple output rays at different angles, preferably with minimal or no translation. Because rays are not translated, the multiple output rays of a single incident ray diverge from a single point. Hence, if multiple output rays pass through a viewer's pupil 110, then they appear to focus at the point from where they are scattered.

[0047] The optical element 106 may comprise one or more holographic optical elements (HOEs), one or more polarisation volume gratings, or one or more surface relief diffractive optical elements, for example. HOEs may be useful in certain scenarios because they can have good reflective efficiency for a design wavelength and incident angle, whist having little interaction with transmitted light.

[0048] Figure 2 shows an arrangement 200 that illustrates the effect of an optical element 202, such as the optical element 106 of the HUD system 100 shown in Figure 1, on a single input ray 204, such as a ray emitted from the image sources 102. As can be seen, the single input ray 204 is scattered, split or otherwise divided into a plurality of bands of output rays (illustrated by the dotted areas), 206, 208, 210 upon interacting with the optical element 202. The optical element 202 is designed to produce the bands 206-210 with properties, such as spacing and / or solid angle, such that scattered rays 206-210 are not incident on both pupils of a viewer simultaneously. The skilled person is aware of many forms of optical element that can be designed to produce these bands, including a HOE, a polarisation volume grating and surface relief diffractive optical elements.

[0049] In this example, the following conditions are met by replications created by the optical element:

[0050] A. The width (illustrated as the distance Xi) of a scattered band, 206-210, at a viewing position is less than the gap between the inside edges of a viewer’s pupils. This is dependent on the viewer’s 212 interpupillary distance and further on the level of pupil dilation. In some examples, Xi may be determined for a particular viewer, or set of viewers. Alternatively, Xi may be determined to encompass as large a population as desired. For example, targeting adult humans, a smallest inter-pupillary distance (IPD) of around 54mm and dilated pupils of around 6mm diameter may be assumed, giving Xi as approximately 48mm. Xi may be reduced further, such as to less than about 48mm to allow for a margin of error in the position of the pupil determined by the pupil-tracking system 112. In some cases, the margin of error in pupil position from the pupil-tracking system may be ±5mm, ±10mm or another value. Where the margin of error is ±10mm, Xi may be smaller than 38mm. B. The gap between bands (illustrated as the distance X2) at the viewing position is greater than the distance between the outside edges of a viewer’s pupils. For a large IPD of around 73mm and dilated pupils of around 6mm, this may be as large as approximately 79mm. Where eye-tracking is used with a margin of error of ±10mm, X2 may be larger than 89mm.

[0051] The result of constraining the bands 206-210 according to conditions A and B is that the viewer 212 will not receive scattered rays from any one of the bands 206-210 through both eyes simultaneously. The conditions may therefore be referred to as “binocular conditions”. It will be understood that Figure 2 depicts Xi and X2 in one dimension, aligned with a line between the centre of the viewer’s pupils. In this example, the extent of the region perpendicular to this (into or out of the page in Figure 2) does not need to be limited.

[0052] Stereoscopic imagery can then be achieved by utilising multiple, independently controllable, bands of additional scattered rays that are configured to fill the gaps between the bands 206-210. An example is illustrated in Figure 3A which shows a diagrammatic arrangement 300 using three independently controllable scattered bands, labelled a’, b’ and c’ . The extent of the bands, a’, b’ and c’ ensure that the eyebox is approximately filled, or tessellated, whilst still ensuring output rays split from a single input ray are not incident on both pupils of a viewer 304 simultaneously.

[0053] In Figure 3A, the scattered outputs corresponding to each input ray, a, b and c, are denoted by dotted, wavy and checkerboard bands respectively. Each of the bands a’, b’, c’ can be independently controlled to ensure that appropriate content is received by each eye of the viewer 304. The bands a’, b’, c’ may be produced by generating the input rays a b, c angularly separated by the same amount as the desired output angular separation, such that that the optical element 302 has the same effect on each input ray a, b, c. In this case, the input rays a, b, c may originate from respective image sources arranged at the angles shown. In other words, the plurality of image sources 102 may comprise three image sources, with each image source configured to generate rays a, b, c that result in the bands a’, b’, c’.

[0054] Alternatively, due to the selectivity of some optical elements 302 to incident angle, such as HOEs, the optical element 302 may be formed as the superposition of holograms for the input rays a, b, c. Forming the optical element 302 as a superposition of holograms for the inputs provides freedom to choose arbitrary input angles for the input rays a, b, c, as pictured in Fig. 3A. Other possible methods for separating the input rays a, b, c include, but are not limited to: using light of different wavelength for each input a, b, c which will result in different scattering angles; spatial multiplexing the inputs a, b, c at the optical element 302; time multiplexing; and polarisation. To satisfy condition A above, the width of bands a, b, c (at a predetermined viewing distance) should be less than approximately 48mm. To satisfy condition B above, the width of bands a, b, c (at the predetermined viewing distance) should be greater than approximately 73 / 2 = 36.5mm. A width of bands a, b, c of approximately 40mm satisfies both conditions, with some margin for error, but the disclosure is not limited to this value.

[0055] Figure 3B shows a front view of the arrangement 300 shown in Figure 3A. In particular, Figure 3B shows the extent of the bands a’, b’, c’ with respect to the viewer’s face. It can be seen that a pupil may overlap two bands (for example b’ and c’), in which case both bands may display content corresponding to that pupil. In such a case the other pupil may be wholly in the remaining band (for example a’). A union of the a’ bands may be referred to as a footprint, or light footprint, for a’ . Similar footprints may be defined for the union of the b’ and c’ bands.

[0056] Note that the viewing position discussed above is understood to be a line or plane where a viewer is expected to view the displayed images. For example, at least one of the viewer’s pupils can be expected to be approximately positioned at the viewing position. The distances Xi and X2 are therefore also dependent on a distance between the viewing position and the optical element 302, and increase linearly with increasing distance. It may therefore be useful to determine the viewing position to a certain degree of accuracy to ensure that the conditions A and B are appropriately satisfied. In some cases the HUD may be physically adjustable to adjust the position of viewing plane relative to a combiner. For example, in a car a longer- legged driver may move the seat back and correspondingly have their head further from the HUD than a shorter-legged driver.

[0057] While three bands are described for Figure 3A and 3B, other examples may use more than three bands, such as four or five. Having a larger number of bands may allow for more eye-tracking error to be allowed. For example, where there is a margin of error of ±10mm, the width of the bands may be less than 38mm, and the gap between adjacent copies of a band than 89mm. In this case, three bands, not two, are needed fill the gap, meaning that there will be four bands in total.

[0058] In an alternative example, the binocular conditions may be amended to allow the bands a’, b’, c’ to at least partially overlap at the viewing position. This example will be discussed in further detail with regards to Figure 11, but it is understood that the arrangement 300 shown in Figures 3A and 3B is but one example of the techniques described herein for stereoscopic HUDs.

[0059] Figure 4 shows a side view of an example optical architecture for the arrangement 300 shown in Figure 3A and 3B. As discussed above, in the example shown in Figure 3A, the input rays a, b, c originate from a respective image source 402, 404, 406 configured to direct rays at the optical element 302 at a desired angle to result in the banding structure a’, b’, c’.

[0060] Each image source 402, 404, 406 projects an image which is focussed on the optical element 4302. Only rays 408 emitted from one of the image sources 404 are shown for simplicity, though rays emitted from the other image sources 402, 406 follow by analogy. The side view shown in Figure 4 illustrates how the vertical nature of the band structures are formed, while Figure 3A illustrates how the repeating horizontal nature of the bands are generated.

[0061] Pupil-tracking in this example allows the HUD system 100 to know which band(s) a’, b’, c’ the viewer’s pupils are in. Some examples may also allow for simulated parallax with head movement, by displaying a view of the content that changes as a viewer’s pupils translate with head movement.

[0062] In the arrangement 300 shown in Figures 3 and 4, multiple output rays that are split from a single input ray diverge from a point on the optical element 302 . While this arrangement is able to produce stereoscopic imagery with a large FoV, the focal depth of the images is fixed at the position of the optical element 302.

[0063] It may be desirable to have the focal depth of the images away from the optical element 302, such as beyond the optical element 302 relative to the viewer. This may reduce focus rivalry and / or vergence-accommodation conflict (VAC). Focus rivalry occurs when the focus depth of the image on the display system differs significantly from the real-world scene it is combined with. Considering an automotive system, drivers typically focus some distance outside the vehicle on the road ahead, so images focussed on the windscreen cannot easily be maintained in focus, creating focus rivalry. VAC is where a perceived depth (from the stereoscopic left and right images) differs from the actual focal depth of the eye. It may be desirable to reduce these effects where possible.

[0064] The example of Figure 5 allows the HUD to be displayed with a focal point different from the optical element 502. A further constraint is applied to ensure that a single pupil does not receive multiple copies of a single input ray within a single band a’, b’, c’. Figure 5 shows this additional constraint on ray geometry by analogy to Figure 2. As can be seen, there is still the binocular constraint on the distances Xi (i.e. the widths of the bands 506, 508, 510) and X2 (i.e. the minimum distance between adjacent bands at the viewing position), but now there is the additional constraint that ensures that a single pupil cannot see multiple copies of a single input ray. In particular, a distance denoted X3, corresponding to a minimum distance between adjacent copies of an input ray within a single band 506-510, at a nominal eyebox distance, is greater than a typical viewer’s pupil size.

[0065] The combined effect of constraining the distances Xi, X2, X3 ensures multiple output rays split from a single input ray are neither incident on a single pupil simultaneously, nor incident on both pupils of the viewer simultaneously (i.e. for any single input ray, a viewer can see at most one output ray). The focal depth of the image perceived by the viewer can be at a different position than the optical element, while also providing a very large FoV. This allows the HUD system to be able to generate images whose focal depth is closer to the environment beyond the combiner 104. This reduces the effects associated with focal rivalry and VAC.

[0066] Figure 6 shows an arrangement 600 illustrating how an individual image point from an image source 602 may be focussed.

[0067] Figure 6 illustrates rays 604 for some central point as well as truncated rays 606, 607 for two other points that are not considered, but which are shown for context. The action of the optical element 606 in the arrangement 600 produces multiple angularly separated bands as has been discussed above However, in this example, the focal point of the generated image is at some point before the optical element 606, not at the optical element 606 as in the examples shown in Figures 2-4. Thus, multiple angularly separated virtual images are produced at some distance beyond the optical element 606. For simplicity, Figure 6 shows these multiple image points approximately at infinity (i.e. pairs of rays are approximately collimated in the optical path between the optical element 606 and the viewer). However, it is understood this is for illustration purposes only, and that the focus may be at some other chosen plane beyond the optical element 606.

[0068] The constraints governing the splitting of rays are similar to those shown in Figure 5, but now an input beam of rays 604 is being considered since the rays do not converge on the optical element 606. In particular, the effect of the optical element 606 on the beam 604 is to produce multiple beams 610, 610’, 612, 612’, 614, 614’ having some width, X4, at the viewing position. As mentioned above, in this example, the focus of the beam 604 is such that the multiple beams 610, 610’, 612, 612’, 614, 614’ are approximately collimated.

[0069] The distance X3 now indicates a spacing between neighbouring pairs of beams 610, 610’; 612, 612’; and 614, 614’, rather than two copies of a single ray. As before then, X3 must be greater than the typical pupil size of a viewer in order for rays from a maximum of one of the neighbouring beams 610, 610’; 612, 612’; and 614, 614’ to reach the viewer’s pupil.

[0070] In a similar manner to the arrangement shown in Figure 2, there are multiple positions where the image point is not visible. By analogy with Figure 3A then, multiple additional, independently controllable, channels may be used to ensure that the eyebox is approximately filled, or tessellated, thereby ensuring that images are always received by the viewer.

[0071] As well as having the independently addressed bands, a’, b’, and c’ (as shown in Figures 3A and 3B), that impose the “binocular conditions”, there may be additional independently addressed types of pupil within each band (for example, a’i, a’2, a’3 within the a’ band). Figure 7 shows an arrangement 700 of beams scattered from an area on the optical element 606 corresponding to nine such independent channels, as sampled at a nominal eyebox location (i.e. a viewing position). For example, if an image source associated with an a’i band displays a single point, illuminating a patch on the optical element 606 (similar to the drawing of Figure 6), then corresponding beams emanating from that patch of the optical element 606 will pass through all of the ‘ai’ areas of Figure 8, and none of the other areas.

[0072] To achieve this, conditions A and B may be imposed on the width of a band, and the gap between bands respectively. Additionally, the following condition may be met:

[0073] C. A gap between multiple copies of a single channel (e.g. the gap between two copies of an ai) need to be greater than the size of a viewer’s pupil. The maximum width of a pupil corresponds to the width when a pupil is dilated. A typical dilated pupil has a width of around 6mm, so that the gap between multiple copies of a single image channel may be greater than around 6mm.

[0074] Each channel, and correspondingly each image source, may be produced by a respective optical module. Alternatively, an optical module may be used to produce two or more of the channels. For example, a single optical module may generate two of the channels via temporal multiplexing at a sufficient speed such that the viewer perceives the outputs of the two channels simultaneously through persistence of vision.

[0075] In the arrangement 700, the a, b, and c bands prevent binocular crosstalk so that left and right eyes are never in the same band, while the 1, 2, 3 subtypes prevent a single pupil from seeing inconsistent copies (i.e. a pupil never spans two copies of a channel). In other words, there are 9 sub-bands, al-c3. The arrangement of sub-bands within the eyebox may itself be determined. For example, the hexagonal arrangement of sub-bands at the eyebox may be selected because such a tessellation may be an optimal arrangement of packing circles. This reduces the regions not addressed by any of the channels. In other examples, other tessellations may be used, such as square grids, to tile the channels.

[0076] Figure 8 shows the properties of the arrangement 700 more clearly, by superimposing circles 802, 804 representing a viewer’s pupils with typical pupil size and IPD. As can be seen, the circle 802 representing the left pupil overlaps the ‘a’ and ‘b’ bands, and correspondingly the circle 804 representing the right pupil is within the ‘c’ band. Similarly, the gap between neighbouring ‘cz’ regions is sufficiently large that the viewer's pupil cannot span two different ‘cz’s. In this case the circle 802 representing the left pupil intersects a2, as and bi regions, and the circle 804 representing the right pupil intersects ci, C2 and C3 regions.

[0077] Figures 7 and 8 show the footprint of the nine different channels at a viewer for a single patch of an optical element (such as the optical element 606 shown in Figure 6). Whilst any patch will look similar, different patches may produce translated footprints. For instance, the viewer may see different regions of an image through different channels. It is understood that there may be more or fewer than nine channels, depending on the characteristics of the viewer, and that the arrangement shown in Figures 7 and 8 is but one example.

[0078] Figure 9 shows an example optical architecture 900 for the arrangements 600, 700 shown in Figures 6 and 7. The optical element 606 may be configured to produce multiple images of an exit pupil of an image source in a nominal eyebox plane 908. In some examples, it may be desirable for the multiple channels within a band to share a single optical module. For instance, it may be useful to use an optical module with independently controllable image sources for three exit pupil regions, which may be time-multiplexed. In this example then, the optical architecture 900 comprises three optical modules 902, 904, 906 configured to generate nine image sources, the ‘a’, ‘b’ and ‘c’ bands respectively, and to generate the ‘1’, ‘2’, ‘3’ subbands via time-multiplexing.

[0079] For simplicity of illustration, Figure 9 shows only rays corresponding to the ‘b’ band, though the ‘a’ and ‘c’ bands are understood to follow by analogy. The ‘b’ image sources, corresponding in this example to the same optical module 904, produces an image in an image plane 910, indicated by a chain-dashed line. This plane may be chosen such that image points appear at the desired focal depth when reflected by the optical element 606, as illustrated.

[0080] As discussed with regards to the arrangements illustrated in Figures 2-4, it may be useful to utilise pupil-tracking to allow the display to 1) know which band(s) (ai-c ) each of the viewer’s pupils are in, for each image point, and also 2) provide simulated parallax with head movement, by displaying a view of the content that matches the user’s eye position.

[0081] As mentioned above, a viewer’s pupil does not have to be, and is indeed extremely unlikely to be, in the same channels for all image points. The display may therefore identify the location of the viewer’s pupils, calculate which channels each image point is visible through, and address content to the correct channels correspondingly.

[0082] Figure 10 shows an example method 1000 of displaying one or more images by a stereoscopic HUD system, such as the stereoscopic HUD system 100 shown in Figure 1. At block 1002, the method 1000 comprises determining positions of left and right pupils of a viewer. The left and right pupil positions may be determined using a pupil-tracking system, such as the pupil-tracking system 110 shown in Figure 1.

[0083] At block 1004, the method 1000 comprises determining left and right images for viewing by the left and right pupils. This may involve determining image views of a stereoscopic image to be displayed at each pupil.

[0084] At block 1006, the method 1000 comprises determining which image points of the left and right images are to be displayed by which image sources The image sources may be two or more of the plurality of image sources 102 shown in Figure 1.

[0085] At block 1008, the method 1000 comprises displaying the image points by the image sources via an optical element. The optical element may be any of the optical elements 106, 202, 302, 502, 606 shown in Figures 1 to 9. The optical element is configured to split an incident ray into multiple output rays at respective different angles. The optical element is further configured such that output rays replicated from a single input ray are not incident on both of the left and right pupil positions at the viewing position. These conditions ensure that the HUD system is capable of generating stereoscopic imagery by preventing cross talk between different image channels at the viewer.

[0086] The method 1000 may be repeated periodically or aperiodically to update the ensure that the appropriate content is being displayed at the correct pupil as the viewer moves. The positions of the left and right pupils and images to be displayed at each pupil may be determined by a controller, such as the controller 114 shown in Figure 1.

[0087] As has been mentioned above, stereoscopy can still be achieved where the bands a’, b’, c’ shown in Figures 3A and 3B are not necessarily restricted to the dimensions described, overlap. An example of such an arrangement 1100 is shown in Figure 11. In this arrangement 1100, a scattered band a’, b’, c’ is wide enough that both eyes may see it at once. This violates condition A described with respect to Figure 3. However, stereoscopic display of images may still be achieved by not displaying any points that would be visible to both eyes into that band, or channel, provided that there is enough redundancy to display these points through other channels instead.

[0088] As can be seen, the bands a’, b’, c’ in Figure 11 are similar to those shown in Figure 3 A, but where the width of each of the bands a’, b’, c’ have been expanded, allowing them to overlap. Because the bands a’, b’, c’ are wider, left and right pupils might now both be in an a’ band, for example. However, in that case if the left eye is in an a’ / b’ overlap region then the right will be in an a’ / c’ overlap region, so a’ may be unused for that image point, while b’ can be used for the left eye and c’ for the right. For instance, a controller, such as the controller 114, may be configured to prevent an image source associated with the a’ band from producing light at that image point.

[0089] This can be extended to the arrangement 700 shown in Figure 7. In this case, now the regions ai, a2, as may be expanded so as to overlap. An example of this type of arrangement 1200 is shown in Figure 12. A pupil may now, for example, span across multiple copies of a single channel, but there will always be at least one channel which it is wholly inside, so that one may be used for that pupil, and the others may be unused. Again, a controller, such as the controller 114, may be configured to control the image sources accordingly. The above embodiments are to be understood as illustrative examples of the invention.

[0090] Further embodiments of the invention are envisaged. It is to be understood that any feature described in relation to any one embodiment may be used alone, or in combination with other features described, and may also be used in combination with one or more features of any other of the embodiments, or any combination of any other of the embodiments. Furthermore, equivalents and modifications not described above may also be employed without departing from the scope of the invention, which is defined in the accompanying claims.

Claims

CLAIMS1. A stereoscopic Head-Up Display, HUD, comprising: a plurality of image sources for generating images; a pupil-tracking system configured to determine left and right pupil positions of a viewer; an optical element positioned in an optical path between the plurality of image sources and a viewing position, wherein the optical element is: configured to split an incident ray into multiple output rays at respective different angles, and configured such that output rays split from a single input ray are not incident on both of the left and right pupil positions at the viewing position; and a controller configured to cause at least one first image source of the plurality of image sources to display at least an image point corresponding to a first image for the left pupil and to cause at least one second image source of the plurality of image sources to display at least an image point corresponding to a second image for the right pupil, the optical element and the configuration of the plurality of image sources combining such that rays of the first image are received at the left pupil position and rays of the second image are received at the right pupil position.

2. The HUD according to any preceding claim, wherein each image comprises an image view of a stereoscopic image.3 The HUD according to any preceding claim, wherein the plurality of image sources comprises three or more image sources.

4. The HUD according to any preceding claim, wherein the optical element is configured such that the multiple output rays split from an incident input ray form multiple discrete bands of output rays.

5. The HUD according to claim 4, wherein the optical element and the image sources are configured such that bands corresponding to rays from each of the image sources are substantially space-filling at the viewing position.

6. The HUD according to any preceding claim 4 or claim 5, wherein a width of a band at the viewing position is less than 50mm.

7. The HUD according to any of claims 4 to 6, wherein a minimum distance between bands corresponding to a single input ray is greater than 75mm.

8. The HUD according to claim 4 or claim 5, wherein bands corresponding to input rays associated with different image sources at least partially overlap.

9. The HUD according to any preceding claim, wherein the optical element is configured such that multiple output rays split from a single input ray are not incident on a single pupil position at the viewing position.

10. The HUD according to claim 9, wherein the first image and the second image have a same focal depth behind the optical element with respect to the viewing position.

11. The HUD according to claim 9 or claim 10 when dependent on claim 4, wherein the optical element and the image sources are configured such that when an image source illuminates a region of the optical element, a plurality of discrete sub-bands of output rays are formed within a band.

12. The HUD according to claim 11, wherein a minimum distance between nearest sub-bands corresponding to a single image source, at the viewing position, is greater than 6mm.

13. The HUD according to any of claims 11 or 12, wherein a union of the sub-bands corresponding to each of the image sources substantially tessellates at the viewing position.

14. The HUD according to any of claims 9 to 12, wherein sub-bands corresponding to different image sources at least partially overlap.

15. The HUD according to any of claims 9 to 14, wherein there are at least nine image sources.

16. The HUD according to any preceding claim, wherein the optical element comprises one or more holographic optical elements.

17. An automotive HUD comprising: a windscreen; and the HUD according to any preceding claim, wherein the optical element is coupled to the windscreen.

18. The Automotive HUD according to claim 17, wherein the optical element comprises a holographic optical element mounted to the windscreen.

19. The automotive HUD according to claim 17 or claim 18, wherein the optical element is configured such that unintended reflections of the image sources are not directed to the viewing position.

20. A method comprising: determining a position of a left pupil of a viewer; determining a position of a right pupil of the viewer; determining a first image for viewing by the left pupil; determining a second image for viewing by the right pupil; determining at least one first image source for displaying at least part of the first image based on the position of the left pupil; determining at least one second image source for displaying at least part of the second image based on the position of the right pupil; and displaying at least an image point corresponding to the first image by the at least one first image source and displaying at least an image point corresponding to the second imageby the at least one second image source via an optical element positioned in an optical path between the first and second image sources and the positions of the left and right pupils, wherein the optical element is: configured to split an incident ray into multiple output rays at respective different angles, and configured such that output rays split from a single input ray are not incident on the positions of the first and second pupils simultaneously.

21. The method according to claim 20, wherein the optical element is configured such that multiple output rays split from a single input ray are not incident on the position of the first pupil or the position of the second pupil.

22. An optical element for a stereoscopic HUD, the optical element comprising a surface feature configured to split an incident ray into a plurality of discrete output rays at different angles, wherein: the angles of the discrete output rays are dependent on a property of the incident ray, and the angles of the discrete output rays are such that the discrete output rays are not incident on the positions of first and second pupils at a viewing position simultaneously.

23. The optical element according to claim 22, wherein the property of the incident ray comprises one or more of an angle of incidence and a wavelength.

24. The optical element according to claim 22 or claim 23, wherein the surface feature comprises a volume holographic grating.

25. The optical element according to claim 22 or claim 23, wherein the surface feature comprises a coating.

26. The optical element according to any of claims 22 to 25, wherein the surface feature is configured such that the discrete output rays form bands having a width at the viewing position that is less than 50mm.

27. The optical element according to claim 26, wherein a minimum distance between adjacent bands is greater than 75mm at the viewing position.

28. The optical element according to any of claims 22 to 27, wherein the angles of the discrete output rays are such that two or more of the discrete output rays are not incident on a single one of the left and right pupil positions simultaneously.

29. A stereoscopic HUD comprising the optical element according to any of claims 22 to 28.

Images