Robotics
Systems

Robotic Product Design / System Integration / Autonomous PerceptionSelected Work / 2024–2025

Overview

Two robotic platforms that connect mechanical development, sensor integration and perception in real-world environments.

The work moves across two scales: shaping and integrating a wheeled humanoid platform, then building a mobile perception system that can inspect inventory through RGB, depth and RFID sensing.

ScopeMechanical designSystem integrationSensor integrationRobotic perception

One practice / Two systems

FROM PHYSICAL ARCHITECTURE
TO ENVIRONMENTAL UNDERSTANDING.

01Embodied platform

Structure, actuation, enclosure development and full-system assembly.

02Robotic perception

Mobility, multi-modal sensing and software that interprets a changing warehouse.

01 / Wheeled humanoid robot

MAKE THE ROBOT READ AS ONE SYSTEM.

The platform combines a wheeled mobile base, dual-arm manipulation, vision hardware and a modular robotic architecture. My contribution focused on translating those technical requirements into a buildable, testable product form.

Completed wheeled humanoid robot — WHEELED HUMANOID PLATFORM / INTEGRATED FORM

01Robot styling and structural layout
02Prototype assembly and evaluation
03Design validation support
04Algorithm and programming support

Prototype evolution

STRUCTURE FIRST.
INTEGRATION NEXT.
FORM LAST.

Early mechanical humanoid prototype with exposed structure — 01Mechanical prototypeEarly structure + actuator integration

Humanoid robot during control and system integration — 02System integrationControl setup + assembly evaluation

Refined humanoid robot with final enclosure — 03Form refinementEnclosure development + full assembly

02 / Autonomous inventory inspection

GIVE A MOBILE ROBOT
A USEFUL VIEW OF THE SHELF.

A vertical sensing module extends the mobile base from floor-level navigation to elevated storage inspection. The system combines visual, depth and identification-based sensing to capture inventory data across stacked warehouse environments.

Autonomous inventory robot with vertical sensing mast — FULL PLATFORM / MOBILE BASE + VERTICAL SENSING MODULE

Close view of the robot camera and sensing module — RGB + DEPTH SENSING

Rear view of the autonomous inventory robot — ON-BOARD COMPUTING + HARDWARE INTEGRATION

System architecture / Perception pipeline

SENSE. ALIGN. COUNT.

The interaction layer connects AGV control with stacked-box counting. RGB and depth acquisition produces point clouds for spatial processing, while RFID and barcode recognition add item-level identification.

Architecture diagram linking AGV control, image and point-cloud acquisition, barcode recognition and box counting — SYSTEM ARCHITECTURE

Robot perception interface showing inventory recognition and point-cloud views — PERCEPTION INTERFACE / IMAGE + POINT CLOUD

Acquire

Capture synchronized RGB, depth and identification data.

Interpret

Preprocess point clouds and filter spatial information.

Identify

Recognize stacked boxes, barcodes and inventory location.

Validate

Test sensor placement and system behaviour on the physical robot.

Outcome

A wheeled humanoid robot progressed from exposed mechanical prototype to an integrated form, while an autonomous inventory platform combined RGB, depth and RFID sensing with point-cloud processing and stacked-box counting.

Across both platforms, product design became a systems discipline: making sensors accessible, structures serviceable and technical behaviour legible enough to evaluate in the real world.

Next project

MORPH

Digital Persona / AI-assisted Persona Design / Social Behaviour

↗

RoboticsSystems