Applications & Implications
What will it do?
My project is a first venture into Social Robotics powered by AI. It uses a Raspberry PI 5, locally hosting an LLM model that inputs prompts with Text to Speech, and outputs the results with Speech To Text. The robot also has a 360 rotating base, allowing it to freely rotate in any direction. This rotating base will enable a person tracking system to detect the current user interacting with with robot, and adjust the base position to always face it.
Who’s done what beforehand?
This video from user Brenpoly on YouTube showcases the step by step process on how the user recreated BMO, a cartoon character from the Adventure Time series, in real life. His project involved the same “text to speech” and “speech to text” functionality I wanted to implement into my project. This video served as a great starting point and inspiration for my final project. The only thing missing in Brenpoly’s design is the mechanical side, as his BMO is stationary, and its more of a “pretty casing” overall. My project involves movement which, whilst still very simple in nature, is a step further in the Social Robotics direction.
What sources will you use?
Some of the sources compiled to this point in the development of the project are, but no limited to:
- Micropython’s official Wiki
- INMP441 microphone modules Datasheet
- MEMS microphone setup tutorial
- 3D printed planetary gear system example
- BOSL2 documentation page
- Raspberry Pi Os image creation tool
- Text to Speech with Wisper tutorial
What will you design?
I will design a “robot” with 4 dimensions: Programming, Mechanics, Electronics, and its casing. The programming dimension involves the AI processing, TTS and STT functionalities. The electronics dimension involves the power supply, PCB and wiring of the robot. The Mechanics dimension involves the planetary gear system on the base of the robot, allowing it a full 360 degrees rotation. The casing dimension involves giving the system a “cute and playful” aspect, in the shape of a cartoon-ish octopus, mascot of Ibero Puebla’s Artificial Intelligence and Augmented Reality Lab, (LIARE).
What materials and components will be used?
Each dimension of the project involves its own set of pieces:
| Programming | Mechanics | Electronics | Casing |
|---|---|---|---|
| Raspberry Pi 5 as central brain unit | NEMA stepper motor to move the rotating base | Custom PCB with the following components: | Custom 3D printed “Body” piece in PLA |
| MEMS Microphone module (INMP441) | Custom 3D printed planetary gear system + base in PLA | drv8825 driver module for the stepper motor | Custom 3D printed “Tentacle” pieces in PLA |
| Xiao ESP32S3 + Sense Module for person tracking | Xiao ESP32C3 + Antenna for stepper motor control | ||
| Bluetooth speaker for the Raspberry Pi 5. | 12v to 5v buck converter to power the ESP32C3 |
Some aspects of the robot listed on the planning page of this final project where scrapped, as including them into the final product involved a complete redesign of the established systems and mechanics, obtaining international imported pieces not available in our Lab, or a high complexity implementation out of scopes for this context.
Where will come from?
Most of the material used for this project is available to us students of Ibero Puebla for free. This include, but are not limited to: Filament, micro-controllers, SMD components (resistors and capacitors mainly), PCB copper plates, soldering equipment, 3D printers, etc. Some material, like the stepper motor and the main power supply unit, where either bought or already owned by me.
How much will they cost?
The following chart lists estimated costs for this project. This lists excludes filament costs:
| Item | Units | Cost per unit (USD) | Cost (USD) |
|---|---|---|---|
| Raspberry Pi 5 | 1 | $85.00 | $85.0 |
| Xiao ESP32C3 | 1 | $4.90 | $4.90 |
| Xiao ESP32S3 Sense | 1 | $13.90 | $13.90 |
| Xiao antenna | 2 | $2.20 | $4.40 |
| NEMA stepper motor | 1 | $13.50 | $13.50 |
| drv8825 driver module | 1 | $2.50 | $2.50 |
| Ball bearing | 4 | $1.25 | $5.00 |
| INMP441 MEMS microphone module | 1 | $2.25 | $2.25 |
| 12v 10A power supply | 1 | $20.00 | $20 |
| 12v to 5v buck converter | 1 | $3.50 | $3.50 |
| Mini Bluetooth speaker | 1 | $5.25 | $5.25 |
| USB to USB-C cable | 1 | $5.25 | $2 |
| Rubber bands | 6 | $0.10 | $0.60 |
| TOTAL | $165.00 |
What parts and systems will be made?
The main systems present on the robot are:
- AI Processing System
- Text to Speech System
- Speech to Text System
- Local LLM Processing
- Person Tracking System
- Planetary Gear System
- Communication Protocols (UDP, Serial)
The following parts where produced, mainly for the mechanical and casing dimensions of the project:
- Planetary Gear System.
- x4 Planet Gears.
- x1 Sun Gear.
- x1 Carrier.
- x1 Outer Ring Gear.
- Supporting Base for the Planetary Gear System.
- Sphere Shaped Main Casing.
- x6 Tentacle Pieces.
- Supporting Base for the Raspberry Pi 5.
- Supporting Base for the Main Casing.
Plus, a custom PCB was fabricated using an XTool F1 Ultra laser machine.
What processes will be used?
The main processes involved in the creation of this project are:
- 3D printing.
- Laser engraving for the PCB.
- Embedded programming (Micropython) and scripting (Python).
- AI assisted research and development.
Complementary processes like laser cutting or vinyl cutting might be used to produce “decorations” for the final result.
What questions need to be answered?
How is AI incorporated into social robots?
How can we give life-like functionality to and AI powered social robot?
How can an AI interact with its “body”?
What are the limitations of processing AI in a Single Board Computer?
Where will the results of this final project take our LIARE lab? For further research and development of AI powered robots
How far can my acquired knowledge during the FAB Academy 2026 take me and my project?
How will it be evaluated?
My personal criteria for evaluation is based on a “pyramid-like” hierarchy of deliverables. The “base”, or the bare minimum requirement for this final project, consists of a functioning, octopus-shaped AI model that users can speak to and receive responses from. The next step is adding a rotating base to introduce motion and give the robot “life”. The final phase involves a person-recognition system that tracks the user’s position and adjusts the rotating base to face them at all times.
Ultimately, the evaluation of this project should judge only the completed deliverables, without penalizing the absence of later phases as a failure to deliver. My area of study is Software Engineering, so everything delivered would be my very first approach at a formal application of 3D design, mechanics, electronics production, etc.