Final Project¶

The Idea¶

What does it do?¶

The delivery robot is an indoor service machine designed to transport and deliver small items from one place to another. It can be remotely controlled through a web interface hosted on the ESP32 and operates on Wi-Fi Access Point mode, allowing users to move the robot without any external network.

The robot can:¶

• Carry small objects inside its storage compartment

• Move forward, backward, left, and right through hoverboard motors

• Detect obstacles using IR sensors

• Provide a simple and efficient way to move items inside labs, workshops, and educational spaces

• Be expanded later with SLAM, lidar scanning, and autonomous navigation using ROS2

Currently, the robot performs manual delivery tasks and demonstrates the full integration of electronics, mechanical design, embedded programming, and digital fabrication.

Who’s done what beforehand?¶

Before developing this project, similar delivery robots were created by various companies and makers. These examples helped define the direction and purpose of my own robot.

Commercial Delivery Robots¶

• Starship Robotics

Small autonomous delivery robots used on university campuses for food and parcel delivery. They use cameras, lidars, GPS, and advanced navigation algorithms.

• Amazon Scout

An autonomous six-wheel delivery vehicle designed for last-mile package delivery. It uses a combination of sensors and AI for navigation in outdoor environments.

• Yandex Rover

A delivery robot used for courier services in Russia, capable of navigating sidewalks and urban environments.

Fab Academy Alumni Projects¶

Several Fab Academy students also created delivery or autonomous robots:

• Autonomous indoor carriers — projects involving ESP32, Raspberry Pi, or custom PCBs to move objects inside buildings.

• Robots with hoverboard motors — many alumni explored hoverboard-based drive systems similar to mine.

• Mobile robots with web interfaces — controlling robots over Wi-Fi is a common approach demonstrated in previous years.

These existing examples showed the feasibility of a small indoor delivery robot and inspired the structure of my own project. However, my robot uses a fully custom PCB, CNC chassis, and a hybrid ESP-IDF + web interface control system, making it unique to the Fab Academy assignment.

Development process¶

What I Designed¶

My final project integrates multiple design components developed during Fab Academy. Each system was created, tested, and documented in different weeks of the course. Below is a summary of the main elements I designed myself.

1. Custom PCB Design and Production¶

I designed a custom ESP32 control board for the robot, which provides:

• UART communication with the hoverboard motor driver

• GPIO expansion for sensors

• Protection circuitry (diodes, capacitors, regulators)

The PCB was fully designed by me, milled and soldered in the lab.

Design files and documentation: View PCB week

2. Robot Chassis (CNC Milled Base)¶

I designed the mechanical frame of the robot using CNC milling. The chassis includes:

• 12 mm plywood structure

All parts were parametrically modelled and cut using a CNC machine.

Documentation: View CNC week

3. 3D-Printed Body (Enclosure)¶

The external shell of the delivery robot was designed in Fusion. It consists of:

• Multi-part PLA shell

• Mounting inserts for screws

• Sensor windows (IR sensors, lidar future support)

• Hinged lid for storage compartment

These parts were printed on Bambu Lab printers.

Documentation: View 3D printing week

4. Embedded Software (ESP-IDF Firmware)¶

I developed custom firmware using ESP-IDF (C/C++) to control the robot:

• Wi-Fi Access Point mode

• Web server for controlling the robot

• UART protocol for hoverboard motors

This firmware runs the whole movement system.

Documentation: View Embedded Programming week

5. Web Interface (Control Panel)¶

I built a lightweight HTML/JavaScript web interface to manually control the robot. Features include:

• Forward / Back / Left / Right / Stop commands

• Works fully offline through ESP32 AP mode

This interface communicates directly with ESP32 over HTTP.

Documentation: View Networking week

Documentation: System integration

Here is code¶

Summary¶

All core parts of the delivery robot — electronics, mechanics, software, and interface — were entirely designed and implemented by me. This project demonstrates the integration of digital fabrication, embedded systems, machining, rapid prototyping, and programming learned throughout Fab Academy.

Final project poster¶

Final project video¶

Final Conclusion¶

My final project was an indoor delivery robot designed to transport small items inside labs and workspaces. I wanted to build a machine that could move objects for students and staff, making daily tasks easier. At first, the idea felt too big, but week by week, using everything I learned during Fab Academy, the robot slowly became real.

For this project, I used:¶

• 3D printing to fabricate the robot’s body parts

• CNC milling to create the wooden chassis

• A custom ESP32 PCB that I designed, milled, soldered, and programmed

• Hoverboard motors controlled through a UART protocol

• A web interface to move the robot over Wi-Fi

• Embedded programming to integrate sensors and motor control

Of course, not everything worked on the first attempt. My first PCB had issues, the robot frame needed redesign, and communication with the hoverboard driver required several tests. But I fixed each problem, learned from every mistake, and kept moving forward.

Through this process, I gained practical knowledge about mechanical structure, electronics design, power management, and communication systems. Every challenge - from motor calibration to PCB production - made the final result stronger.

The best moment was when the robot successfully moved using my own custom board and web interface. It showed me that I can take a complex idea and turn it into a functioning machine built entirely with the skills I learned in Fab Academy.

In the future, I plan to upgrade my delivery robot with Jetson Nano, a LiDAR sensor, and an Intel RealSense depth camera, all running on ROS2. These additions will allow me to implement autonomous navigation, SLAM, and advanced perception features in the next version of the robot.

During Fab Academy, I learned how to:¶

• Design, mill, and assemble electronic boards

• Build reliable mechanical parts

• Program microcontrollers and interfaces

• Combine hardware, software, and fabrication into one complete system

• And most importantly — how to solve problems independently

I am very grateful to my instructors and to the Waag and Dilijan Fab Lab. This course helped me grow as an engineer and gave me confidence to work on even more advanced robotics and embedded systems projects. Now I feel ready to continue learning, improving, and building the next generation of intelligent machines.

Final project files:¶

Download file for 3D printing

Chassis file

Kicad files

Gerber , drill , gcode and png files

Final arduino code