FILE: 2570_CFB
Distributed Masonry
Corrugated Brick Assemblies Through Projection-Based Human-Robot Collaboration

Team:

Samet Yılmaz
Jutang Gao
Yilin Lin


Computational Fabrication | 202526 Fall

Princeton University



Instructor

Prof.Dr. Arash Adel
Teaching Assist. Daniel Ruan










FILE: 2570_CFB
Distributed Masonry
Corrugated Brick Assemblies Through Projection-Based Human-Robot Collaboration

Team: 
Samet Yılmaz
Jutang Gao
Yilin Lin


Computational Fabrication | 2025’26 Fall

Princeton University

Instructor

Prof.Dr. Arash Adel
Teaching Assist. Daniel Ruan

Developed in Princeton University’s Computational Fabrication course, this team project explores how computational fabrication reshapes construction through new human–robot labor divisions, integrated production/assembly, and hybrid media. It asks how robotic precision expands bricklaying design space: enabling efficient structural configurations, parameterizing them, and translating them into buildable systems. We built a full-scale wall prototype using Rhino/Grasshopper, a robotic arm, projection guidance, and on-site human fabrication.


01.  Corrugated brick assemblies through projection-based human-robot collaboration




02. Combining two types of running bonds


Input sequence 1: Boolean pattern for brick orientation.

Input sequence 2: Brick relative location to the center line.

Input sequence 3: Horizontal movements of layers.

All three patterns are made repetitive across the entire wall.

03.  Combining two types of running bonds by human



Notes: 

Sequence of laying down bricks in different orientations should be paid attention to, as improper sequence will cause collision. 

Certain amount of gap/tolerance is expected to allow geometric flexibility and accommodate for construction errors.



04. Using NURBS surfaces as input


05. Using NURBS surfaces as input




06. Geometry-based reinforcement



curved           corrugated
07. Geometry-based reinforcement (Corrugated Form)

08. Flaws of this model


F1:
Vertical gaps across more than one course need to be avoided. 

F2: Too much space near edges are used for aligning
09. Flaws of this model


Goal: to even-out brick bonds across the entire structure and centralize bricks for safer placement during fabrication. 

For each iteration: 

For each brick that is not first or last brick in a course, do: 

01. get their overlapped areas with previous course bricks and next course bricks.

02. calculate the differences between current brick center and center of gaps, as translation vectors.

03. sum up weighted vectors (based on which surface we want to prioritize and desired step size for each iteration), and get a potential new location.

04. check if there will be collision with neighbors, if so instead move brick only in surface normal direction.



10. Bonding optimization


11. Assembly to Build
11. Assembly to Build
11. Assembly to Build
12.  Input Surface
12. Generation
12. Optimization
12. Output Bricks



13. Completed  wall

14.  Wall module generation
Lorem 15.  Robotic Arm  Simulation...
15.  Robotic Arm  Simulation
15.  Robotic Arm  Simulation...
15.  Robotic Arm  Simulation
15. Robotic arm simulation
16. Brick assemblies through projection-based human-robot collaboration
17.  Moving TCP to projection frame, projector displaying the desired brick position
18.  Moving TCP to projection frame, projector displaying the desired brick position
19.  Moving TCP to projection frame, projector displaying the desired brick position
20. Brick assemblies through projection-based human-robot collaboration
21.  Mock-up of brick wall
22.  Gallery layout generation
23.  Gallery plan
24.  Gallery section
25.  Interior of gallery 
26.  Exterior of gallery
 Samet Yılmaz © 2024, created by Samet Yılmaz, all rights reserved.

Developed in Princeton University’s Computational Fabrication course, this team project explores how computational fabrication reshapes construction through new human–robot labor divisions, integrated production/assembly, and hybrid media. It asks how robotic precision expands bricklaying design space: enabling efficient structural configurations, parameterizing them, and translating them into buildable systems. We built a full-scale wall prototype using Rhino/Grasshopper, a robotic arm, projection guidance, and on-site human fabrication.