Tile-Based Image Generator for HUDs Reducing Latency

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Head-up displays (HUDs) and helmet-mounted displays (HMDs) face challenges in reducing latency and certification complexity due to the power-hungry and short-lived nature of graphics processing units (GPUs), which lead to obsolescence and high re-certification costs, especially when a significant portion of the image is displayed as a uniform background color.

Innovation Solution

An image generator that divides the image area into tiles, storing only tiles with non-background color changes in memory, using a cache and buffer indices to minimize memory access and reduce latency, and employing a software-implemented interface with FPGA-based rendering to simplify certification and reduce power consumption.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If a graphics processor unit (GPU) is used for image generation, then good performance is provided, but power consumption increases and device lifespan decreases

Engineering Contradiction:
Improveimage generation performanceVSAvoidpower consumption
Core Design Contradiction:
ProductivityVSUse of energy by moving object

Solution Approach 1:

The patent extracts only the essential image generation functionality from a full GPU system, implementing a minimalistic renderer that handles only the critical rendering tasks required for HUD/HMD displays. This eliminates the need for power-hungry GPU components while maintaining adequate performance for the specific application.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent applies local quality by optimizing the rendering system specifically for the uniform background color characteristics of HUD/HMD displays. Instead of using a general-purpose GPU optimized for diverse graphics workloads, the system is tailored to handle the specific pattern of uniform background regions and foreground symbology, reducing overall power consumption while maintaining image generation performance.

Inventive Principle:
Principle #3Local quality

2Productivity

If a graphics processor unit (GPU) is used for image generation, then good performance is provided, but device lifespan decreases

Engineering Contradiction:
Improveimage generation performanceVSAvoiddevice lifespan
Core Design Contradiction:
ProductivityVSDuration of action of stationary object

Solution Approach 1:

The patent extracts only the essential image generation functionality from a full GPU system, implementing a minimalistic renderer that handles only the critical rendering tasks required for HUD/HMD displays. This eliminates the need for power-hungry GPU components while maintaining adequate performance for the specific application.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent avoids using expensive, short-lived GPU components by implementing a simpler, more robust rendering architecture using FPGAs and custom logic that is designed for long-term reliability in avionic environments, matching the 25+ year service life requirements of aircraft systems.

Inventive Principle:
Principle #27Cheap short-living objects (Disposable)

3Productivity

If a graphics processor unit (GPU) is used for image generation, then good performance is provided, but certification complexity increases

Engineering Contradiction:
Improveimage generation performanceVSAvoidcertification complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent extracts only the essential image generation functionality from a full GPU system, implementing a minimalistic renderer that handles only the critical rendering tasks required for HUD/HMD displays. This eliminates the need for power-hungry GPU components while maintaining adequate performance for the specific application.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent changes the fundamental parameters of the rendering system from commercial GPU architecture to FPGA-based custom logic, which allows for deterministic behavior and easier certification to avionic standards such as DO-254 and DO-178C, while still achieving the required image generation performance for safety-critical displays.

Inventive Principle:
Principle #35Parameter changes

4Reliability

If memory data for all tiles is stored, then complete image data is available, but memory bandwidth usage increases

Engineering Contradiction:
Improveimage data completenessVSAvoidmemory bandwidth usage
Core Design Contradiction:
ReliabilityVSLoss of energy

Solution Approach 1:

The patent extracts only the essential pixels from each tile for storage in memory, specifically storing only non-background pixels and their coordinates. This selective extraction approach maintains image data completeness for rendering while dramatically reducing memory bandwidth requirements compared to storing complete tile data.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent applies partial action by storing only the necessary portion of tile data (non-background pixels) rather than complete tile data. This partial storage approach is sufficient for reconstructing the image during rendering while significantly reducing memory bandwidth consumption, especially for tiles dominated by uniform background colors.

Inventive Principle:
Principle #16Partial or excessive action

5Adaptability or versatility

If image data is frequently accessed from memory, then rendering flexibility is maintained, but latency increases

Engineering Contradiction:
Improverendering flexibilityVSAvoidrendering latency
Core Design Contradiction:
Adaptability or versatilityVSLoss of time

Solution Approach 1:

The patent segments the image area into multiple tiles, allowing independent processing and rendering of each tile. This segmentation enables the system to process only the tiles containing non-background pixels, reducing overall rendering latency while maintaining the flexibility to handle different display scenarios by adjusting tile processing as needed.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary identification of non-background pixels within each tile before committing to full tile rendering or memory access. This preliminary action allows the system to skip processing for tiles that contain only background pixels, significantly reducing rendering latency while maintaining flexibility for tiles that do require rendering.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS10565753B2Method and apparatus for generating an image
Publication Date: 2020.02.18 BAE SYSTEMS PLC
  • US10565753B2 patent drawing
  • US10565753B2 patent drawing
  • US10565753B2 patent drawing

AI summary

A method and apparatus are provided for generating image data for output to a display device such as an HUD or HMD defining pixels of an image for display. The image data for each pixel defines a predetermined background colour or a non-background colour. In an example implementation, the image generator comprises at least one image rendering module and a memory accessible by the at least one rendering module providing at least one frame buffer in which image data may be assembled. The at least one rendering module is arranged to process or to generate image data using a tile representation of an image for display, each tile comprising image data defining each of a predetermined number of pixels of the image. Furthermore, the at least one rendering module is arranged to store image data for the pixels of a tile in the at least one frame buffer only in the event that one or more pixels of the tile have respective image data defining a non-background colour or are enabled for display with a non-background colour, the pixels in the tile being otherwise assumed to be intended for display with the predetermined background colour. In this way the need to transfer data to and from the memory when generating and assembling image data is reduced, so reducing image rendering time.