David Du
Why this project?
Building an embedded product with a screen often starts in a very different place from where the engineering ends. The starting point may be a sketch on paper, a Figma screen, a photo of an existing product, a Raspberry Pi or SoM prototype, or simply an idea for what the product should do.
Turning that into a real embedded product requires decisions about the display, MCU/MPU/SoM, LVGL/Qt, memory, interfaces, boot time, lifecycle, volume and production cost. OpenHMI is an experiment in connecting these steps.
Idea / Sketch / Figma → Requirements → HMI Definition → GUI → Embedded Hardware → Working Prototype → Production Architecture
What we are testing
The interesting part isn't simply getting another LVGL demo running. We want to test whether an incomplete product idea can be turned into a useful engineering definition.
For example, “I want a touchscreen controller for a machine” may become a set of requirements: a 7-inch display, capacitive touch, CAN + Ethernet, Linux or RTOS, GUI performance requirements, a boot-time target, enclosure constraints, annual volume, lifecycle requirements and a production cost target. These are questions to resolve, not assumptions to lock in.
Principle: a current architecture is not always a product requirement
A Raspberry Pi, development board or SoM may be perfect for a prototype. When the product moves toward production, we should evaluate whether to keep the SoM, use a more optimized module, or design a custom SoC board. The choice should follow the product's requirements and engineering tradeoffs.
First case: Figma → Embedded Hardware
Project #1 starts from a UI definition rather than a hardware platform. The experiment is to extract screen structure, layout, navigation and visual assets; identify the missing engineering requirements; and bring the interface onto real embedded hardware.
The planned path is toward LVGL/UIBuilder and an ArtInChip D133 EVB. This first case will document the decisions and gaps along that path; a working result is still to be demonstrated.
Have an HMI idea we can test?
I'm looking for a few real HMI examples to test this workflow. You don't need a complete specification. A sketch, Figma design, photo, block diagram or existing prototype is enough.
If the project is a good fit, we'll first help define the HMI architecture and may build a small proof-of-concept on real embedded hardware.