KP.Keaghan Paterson home

Robotics case study

Dexterity, engineered.

Robotic hands and forearms designed to juggle Rubik’s cubes while solving them.

Scroll to explore

01 / 07

Project overview

One system. Many hard problems.

The goal is to coordinate mechanical dexterity, sensing, control, and perception in a single robotic demonstration. Detailed claims and results will be added only after they are verified.

01

Build a full-stack robotics platform

Mechanical design, ECE, controls, and embodied AI in one system

02

Handle objects in unpredictable positions

Perceive, reach, and manipulate parts wherever they land

03

Showcase the system in a memorable demo

Turn the full technical stack into a clear, engaging result

Solve time

To be measured

Juggling duration

To be measured

Success rate

To be measured

Degrees of freedom

To be confirmed

System weight

To be measured

Build cost

To be calculated

02 / 07

Prototype development

Progress, measured in iterations.

Generation 01

Initial CAD concept

Problem discovered
Concept assumptions to validate
Design change
Geometry exploration
Why it changed
Establish a buildable direction

03 / 07

Mechanical design

Built around motion.

This section will explain how the mechanism turns actuator input into controlled hand and forearm movement—from load paths to assembly details.

01

Degrees of freedom

Project-specific detail will be added after design documentation is reviewed.

02

Actuation

Project-specific detail will be added after design documentation is reviewed.

03

Finger mechanisms

Project-specific detail will be added after design documentation is reviewed.

04

Wrist + forearm

Project-specific detail will be added after design documentation is reviewed.

05

Materials

Project-specific detail will be added after design documentation is reviewed.

06

Bearings + joints

Project-specific detail will be added after design documentation is reviewed.

07

Cable routing

Project-specific detail will be added after design documentation is reviewed.

08

Torque + loads

Project-specific detail will be added after design documentation is reviewed.

09

Design for assembly

Project-specific detail will be added after design documentation is reviewed.

10

Mechanical testing

Project-specific detail will be added after design documentation is reviewed.

04 / 07

Electrical + control

Power, signals, decisions.

A clear path from command to motion, with the power distribution, controllers, sensing, communication, and safety layers documented together.

System architecture / placeholder

01

Power

02

Control

03

Actuation

04

Sensing

Replace this map with the final electrical and control architecture diagram.

01

Power system

Project-specific detail will be added after design documentation is reviewed.

02

Controllers

Project-specific detail will be added after design documentation is reviewed.

03

Motor drivers

Project-specific detail will be added after design documentation is reviewed.

04

Sensors

Project-specific detail will be added after design documentation is reviewed.

05

Communications

Project-specific detail will be added after design documentation is reviewed.

06

Wiring + connectors

Project-specific detail will be added after design documentation is reviewed.

07

Safety + faults

Project-specific detail will be added after design documentation is reviewed.

05 / 07

Software + vision

Seeing the cube. Planning the move.

The software story will connect cameras, calibration, perception, motion planning, feedback control, and any machine-learning components without treating them as a black box.

Landscape solve video / placeholder
CAM 01 video / placeholder
CAM 02 video / placeholder
Cube A geometry / placeholder
Cube B geometry / placeholder
Cube C geometry / placeholder

Camera recognition frame / placeholder

Cube · 98%Face · 96%Grip · 91%

Machine-learning identification

Locate, classify, and track.

This panel will explain how camera frames are annotated, how cube faces and grasp points are identified, and how confidence is passed into the motion system.

Live cube-state model / placeholder

Current stateTracking
Remaining moves12
AlgorithmSolver output / placeholder

This view will update as each move is completed, showing the stored geometry, planned sequence, and live solve status.

Recorded camera view / placeholder

CAM 01 — footage pending

Data visualization / placeholder

01

Software architecture

Project-specific detail will be added after design documentation is reviewed.

02

Camera arrangement

Project-specific detail will be added after design documentation is reviewed.

03

Calibration

Project-specific detail will be added after design documentation is reviewed.

04

Object tracking

Project-specific detail will be added after design documentation is reviewed.

05

Cube-state recognition

Project-specific detail will be added after design documentation is reviewed.

06

Motion planning

Project-specific detail will be added after design documentation is reviewed.

07

Feedback control

Project-specific detail will be added after design documentation is reviewed.

08

Machine learning

Project-specific detail will be added after design documentation is reviewed.

09

User interface

Project-specific detail will be added after design documentation is reviewed.

10

Recorded camera views

Project-specific detail will be added after design documentation is reviewed.

11

Code + data visualizations

Project-specific detail will be added after design documentation is reviewed.

06 / 07

Lessons + next steps

The useful part is what changed.

A candid engineering retrospective: decisions that held up, approaches that did not, and the clearest path to a better next version.

01

What worked

Add the strongest design choices and the evidence that supported them.

02

What failed

Document the unsuccessful approaches and what each one revealed.

03

Most important lessons

Summarize the engineering principles that changed the final design.

04

Current limitations

Describe known constraints without overstating system performance.

05

Improvements

Prioritize the changes that would most improve the next iteration.

06

Future goals

Define the next technical milestones for the project.

07 / 07

Full technical deep-dive

The complete design story.

A planned 30-minute walkthrough for mechanical-design hiring managers, technical recruiters, and engineers—focused on the decisions behind the result.