Efficient holographic solutions for vehicle head-up displays

Imagine driving down a busy highway. You need to check your speed and navigation, but glancing down at the dashboard takes your eyes off the road for a critical second. This is where head-up displays (HUDs) come in, projecting information directly onto the windshield. However, current HUD technologies are often bulky and limited to displaying flat, 2D images at a fixed distance, forcing your eyes to constantly refocus between the data and the road.

To address these limitations, a research team led by Professor Qingqing Cheng from the University of Shanghai for Science and Technology and Professor Kun Huang from the University of Science and Technology of China has proposed a new holographic display approach. Their report is published in Advanced Photonics Nexus.

The Challenge: Breaking the Digital Limits

A long-standing challenge in the field is achieving high-quality holographic imaging—using light diffraction to create virtual objects that appear to sit at different depths in the real world (e.g., navigation arrows that seem to lie directly on the pavement). However, calculating these holograms is traditionally computationally expensive.

Conventional methods, typically based on the fast Fourier transform (FFT), are rigid. They generally require the digital image source and the projected image to have matched sampling densities. When scientists try to project a small image (from a display chip) to a large area (the windshield), traditional mathematics requires adding a massive amount of "zero-padding"—essentially empty data—to make the calculation work. This wastes computer memory and slows down the process, making it impractical for real-time vehicle use.

The Solution: "Zoom Lens" for Calculations

The researchers introduced a novel calculation approach: the matrix multiplication (MM) based diffraction method. Instead of using the rigid FFT framework, they restructured the Fresnel integral into a series of flexible matrix operations.

Think of this as a "zoom lens" for calculations. The MM method allows the computer to calculate the hologram for the display chip and the windshield independently. It eliminates the need for zero-padding, effectively removing the redundant computation. In their benchmarks, this method reduced calculation time by approximately 58 percent and significantly lowered memory usage compared to traditional methods.

To validate this approach, the team built a prototype HUD system using the new algorithm. They successfully projected three different virtual images simultaneously at three distinct distances: 0.1 meters, 0.5 meters, and 1.5 meters. In their demonstration, holographic driving information was projected to align perfectly with real-world reference objects such as a traffic cone and a construction worker positioned at different depths, creating a seamless mix of virtual and physical reality.

Crucially, this method can handle extreme size differences—projecting a tiny image and a massive image at the same time—and works for both near-field and far-field displays within a single computational framework.

This technology represents a significant step toward compact, wide-field-of-view AR-HUDs. By making the computational process faster and more flexible, it opens the door for intelligent vehicles that can overlay vital safety alerts directly onto the physical environment without bulky hardware. As the team continues to refine the color and refresh rates, the smart windshield of the future is coming clearly into focus.

For details, see the original Gold Open Access article by X. Sheng et al., “Cross-scale multiplane holograms with decoupled input–output sampling for vehicle display,” Adv. Photon. Nexus 5(1) 016005 (2025), doi: 10.1117/1.APN.5.1.016005