显示性能进展：AR、VR、QLED、OLED

The emerging augmented reality (AR) and virtual reality (VR) applications continue to drive the developmentof the near-eye display (NED) or head-mounted display (HMD) techniques. Various research and development efforts are being made to enhance the traditionalperformance factors of the AR and VR NEDs, suchas field of view (FOV) and angular resolution. Therecent researches, however, also focused on various features enabling realistic and comfortable image presentation, including vergence-accommodation conflict (VAC)mitigation, hard-edge occlusion, and vision correction.Table 1 shows the features of the recently reported orcommercialized NEDs.

The FOV of a NED is the angular size of the displayedvirtual image. Generally, a large FOV is desirable to coverthe FOV of the human visual system, which reaches 160degrees when the eye rolling is considered [15]. Althoughthe maximum FOV of the commercialized VR NED isabout 170 degrees, which is achieved by using doubledisplay devices [11], the typical FOV of the commercialized VR NED is limited to 110 degrees [16]. The effortto enhance the FOV for the VR NEDs is usually focusedon the development of new lens optics [17,18]. In the case of the AR NEDs, the FOV is further limited dueto the requirement of the transparent image combiner.The typical FOV of the commercially available AR NEDsis about 50 degrees [10,14]. The recently reported workfor enhancing the FOV of the AR NEDs included theuse of polarization-dependent grating [19,20]. The use ofthe geometric-phase (GP) lens has also been reported toachieve over 80 degrees FOV by enabling transmissiontype configuration [9,21].

The angular resolution of a NED is defined by thenumber of pixels in a unit degree. The direct approach toenhancing the angular resolution is to increase the pixeldensity of the display panel. The trade-off relationshipbetween the FOV and the angular resolution, however,makes it difficult to achieve a wide FOV and a high angular resolution simultaneously at a given pixel density ofthe display panel [22]. A notable research work in 2019reported the foveated displays. Motivated by the different angular resolution of the human visual system inthe central vision (around 60 pixels per degree) and theperipheral vision (around 30 pixels per degree) [23], thefoveated displays present high-angular-resolution imagesonly within the eye gaze area while maintaining a lowangular resolution in the peripheral area, reducing thetotal system resolution requirement. Several techniqueshave been reported in 2019 to achieve the foveated imagepresentation and dynamic change of the foveated areaaccording to the tracked eye gaze direction [13,24–26],which are summarized in Table 2.

Fig1

Along with the research on the fundamental performancefactors, including the FOV, angular resolution, and VACmitigation, various techniques enabling more realisticand comfortable viewing have also been reported in 2019.One example is the hard-edge occlusion of real objects by the displayed virtual images in the AR NEDs. The usualAR NEDs only add the light for the virtual images to thelight coming from real objects without proper blocking,which makes the displayed virtual images translucent.The hard-edge occlusion techniques block the light fromreal objects at the virtual image position, enabling morerealistic image presentation and enhancing the imagecontrast. Although the first high-edge occlusion technique has been demonstrated nearly two decades ago[37], the recent works in 2019 report a more compactform factor [38] and variable distance occlusion [39].Another example is the AR NEDs with vision prescription lenses [40]. For the widespread use of the AR and VRNEDs, comfortable wearing is important. The compactAR NEDs with vision prescription lenses make it easy towear the AR NEDs on top of the vision-correcting glasses,making the AR NED experience comfortable and morepractical. A subtractive AR NED has also been reportedin 2019 [41]. While most AR NEDs work in additivemode, the subtractive AR NED presents images by subtracting the virtual image portion from the incoming realobject light, making the virtual images more apparentin bright ambient light conditions. Table 4 summarizesthese works.