metaverseDubai/docs/PCG_WORKFLOW.md

PCG Workflow - Dubai Metaverse

Overview

This document describes the Procedural Content Generation (PCG) workflow for the Dubai Metaverse project, including building placement, road props, and vegetation.

PCG Framework Basics

What is PCG?

Procedural Content Generation Framework is Unreal Engine 5's built-in system for procedurally placing and generating content in levels. It uses graph-based workflows to define rules for content placement.

Key Concepts

• PCG Graph: Visual graph defining placement rules
• Input: Splines, volumes, points, or other data
• Output: Placed actors, meshes, or other content
• Rules: Logic defining how content is placed

---

Building Placement PCG

Graph: PCG_BuildingPlacement

Purpose: Procedurally place buildings in the Dubai Marina district

Input Data

• Building Footprints: From OpenStreetMap data or manual placement
• Spline Points: Road network splines
• Volume: District boundary volume

Placement Rules

1. Distance from Roads: Buildings placed 5-10m from road edges
2. Building Spacing: Minimum 2m spacing between buildings
3. Building Types:
   • Residential towers: Near Marina
   • Commercial: Along main roads
   • Background: Fill remaining areas
4. Orientation: Buildings face roads or Marina

Output

• Static Mesh Actors: Building meshes placed in level
• Material Instances: Assigned based on building type
• LODs: Automatically assigned based on distance

Implementation Steps

1. Create PCG Graph:

   • Create new PCG graph asset: PCG_BuildingPlacement.pcg
   • Add input node (spline or points)

2. Define Rules:

   • Add "Surface Sampler" node to sample building locations
   • Add "Point Filter" to filter valid placement points
   • Add "Spawn Actor" node to place building meshes

3. Configure Parameters:

   • Building density
   • Building types
   • Placement constraints

4. Test and Iterate:

   • Execute graph
   • Review placement
   • Adjust rules as needed

---

Road Props PCG

Graph: PCG_RoadProps

Purpose: Place road props (lamp posts, barriers, signs) along roads

Input Data

• Road Splines: Road network splines
• Prop Types: Lamp posts, barriers, signs, etc.

Placement Rules

1. Lamp Posts:

   • Spacing: 20-30m along roads
   • Placement: Road edges, alternating sides
   • Height: Match road level

2. Barriers:

   • Placement: Road medians, pedestrian areas
   • Spacing: Continuous or 2-5m gaps

3. Signs:

   • Placement: Road intersections, important locations
   • Orientation: Face traffic direction

Output

• Static Mesh Actors: Prop meshes placed along roads
• Variations: Random prop variations for variety

Implementation Steps

1. Create PCG Graph: PCG_RoadProps.pcg
2. Add Spline Sampler: Sample points along road splines
3. Add Spawn Actor: Place prop meshes at sampled points
4. Add Variation: Random prop selection and rotation

---

Vegetation PCG

Graph: PCG_Vegetation

Purpose: Place trees, shrubs, and vegetation in the district

Input Data

• Terrain: Landscape/terrain data
• Exclusion Zones: Areas where vegetation shouldn't be placed (roads, buildings)

Placement Rules

1. Palm Trees:

   • Placement: Marina promenade, parks, green spaces
   • Spacing: 5-10m between trees
   • Clustering: Group in natural clusters

2. Shrubs:

   • Placement: Around buildings, along paths
   • Density: Medium to high
   • Randomization: Vary size and rotation

3. Grass:

   • Placement: Open areas, parks
   • Density: High
   • Use foliage system for performance

Output

• Static Mesh Actors: Tree and shrub meshes
• Foliage Instances: Grass and small vegetation
• Variations: Random size, rotation, and type

Implementation Steps

1. Create PCG Graph: PCG_Vegetation.pcg
2. Add Surface Sampler: Sample terrain surface
3. Add Exclusion Filter: Exclude roads and buildings
4. Add Spawn Actor: Place vegetation meshes
5. Add Variation: Randomize size and rotation

---

Traffic PCG

Graph: PCG_Traffic

Purpose: Generate AI traffic on roads

Input Data

• Road Splines: Road network for traffic paths
• Vehicle Types: Car, SUV, supercar variants

Placement Rules

1. Vehicle Spawning:

   • Spawn points: Road intersections, entry points
   • Density: Moderate (not too crowded)
   • Types: Mix of vehicle types

2. Traffic Flow:

   • Direction: Follow road direction
   • Speed: Vary by vehicle type
   • Behavior: Use Chaos Vehicle AI

Output

• Vehicle Actors: Spawned vehicles on roads
• AI Controllers: Vehicle AI for movement
• Traffic Flow: Realistic traffic patterns

Implementation Steps

1. Create PCG Graph: PCG_Traffic.pcg
2. Add Spline Sampler: Sample road splines
3. Add Spawn Actor: Spawn vehicle blueprints
4. Configure AI: Set up vehicle AI controllers

---

PCG Graph Organization

Folder Structure

Content/PCG/
├── PCG_BuildingPlacement.pcg
├── PCG_RoadProps.pcg
├── PCG_Vegetation.pcg
├── PCG_Traffic.pcg
└── SubGraphs/
    ├── PCG_BuildingRules.pcg
    └── PCG_PropRules.pcg

Naming Convention

• Format: PCG_SystemName_Purpose.pcg
• Examples:
  • PCG_BuildingPlacement.pcg
  • PCG_RoadProps_LampPosts.pcg

---

Best Practices

Performance

1. Baking: Bake PCG output to static meshes for performance
2. LODs: Use LOD system for distant content
3. Culling: Use distance culling for props
4. Instancing: Use instanced static meshes where possible

Quality

1. Variation: Add randomization for natural look
2. Rules: Define clear placement rules
3. Validation: Validate output quality
4. Iteration: Iterate on rules until satisfied

Workflow

1. Start Simple: Begin with basic placement rules
2. Iterate: Refine rules based on output
3. Test: Test in level regularly
4. Document: Document rules and parameters

---

Troubleshooting

Common Issues

Issue: Buildings overlapping

• Solution: Add minimum distance filter

Issue: Props not placing correctly

• Solution: Check spline direction and sampling

Issue: Performance issues

• Solution: Bake to static meshes, use LODs

Issue: Vegetation in wrong areas

• Solution: Add exclusion zones/filters

---

Validation

PCG Output Validation

Use scripts/pcg_validation.py to validate:

• Placement rules are followed
• No overlapping content
• Performance is acceptable
• Visual quality meets standards

---

Version: 1.0 Last Updated: [Current Date]