Back to portfolio

Roadside assets & fragment architecture

A reusable 3D roadside library and demo composition model so automotive and partner demos stay consistent without custom props every time.

Context

Navigation and automotive demos need believable roadside detail: barriers, signage, vegetation, and small infrastructure. One-off models per demo do not scale, and partners expect visuals that sit cleanly inside Mapbox Standard and real road data.

Problem

Requests arrived ad hoc; visuals drifted between teams. There was no shared catalog, no agreed style bar against real-world references, and no structured way to assemble scenes for hd-roads and similar stacks without reinventing layout each time.

Approach

Regional HD research across US, EU, and Japan to ground proportions and density. A growing library of 400+ roadside models with naming and reuse rules, plus a fragment-based architecture so demos compose from reusable pieces instead of monolithic scenes.

Style alignment with partner expectations (including BMW-oriented specs where relevant), and filtering in hd-roads so only props that match real road inventory show up for a given context.

Outcome

Fewer bespoke requests, faster demo assembly, and org-wide reuse. Partner-facing demos read as one system rather than a sequence of one-offs.

What next

Expand regional coverage and keep the library aligned as Standard and road data products evolve.