Tommy

Tommy is an Arduino-powered robot that's seriously impressive; he's engineered to autonomously sweep up debris and clear away waste.

This project addresses the growing issue of street littering in Hong Kong by developing an autonomous, large-scale robotic vacuum unit. Designed for off-peak operations, the robot utilizes a combination of ultrasonic sensing for obstacle avoidance and a dual-brush sweeping mechanism to manage refuse in public thoroughfares.

Reflection

This project highlights the necessity of Agile Development. When hardware was delayed, I prioritized software simulation and algorithmic logic. The transition from a basic obstacle-avoider to a robot using the Pledge Algorithm represents a significant step up in embedded systems programming and spatial logic.

Mar 25, 2026: Custom Mechanical Design

Process: Moved beyond off-the-shelf parts to custom engineering.

Tackle: Designed a bespoke Servo Mount for the Sweeper using Onshape CAD software to ensure the 360-degree servos were angled correctly for maximum debris collection.

Mar 11 - Mar 18, 2026: Navigation Logic (The "Pledge" Algorithm)

Process: Implemented the Pledge Algorithm for wall-following and obstacle navigation. Integrated a physical button to toggle the system and an LED indicator for obstacle detection.

Challenge: Simple "Stop-and-Turn" logic resulted in the robot getting stuck in corners or infinite loops.

Tackle: Developed a complex decision-making script that uses a turnCounter variable to track the robot's orientation relative to a wall, allowing it to navigate around obstacles and return to its original heading.

Feb 14 - Feb 28, 2026: Hardware Assembly & Simulation

Process: Finalized the 4-wheel smart car chassis assembly and mounted the Arduino. Used a digital simulation tool (Cirkit Designer) to model the motor driver circuitry before physical wiring.

Challenge: The L9110S motor controller went missing during assembly.

Tackle: Substituted the L9110S with a breadboard-based L298N simulation to verify PWM (Pulse Width Modulation) signals, ensuring the sweeping brushes could operate at variable speeds once the hardware was replaced.

Jan 24 - Jan 31, 2026: Prototype Planning & POC

Process: Drafted a Bill of Materials (BOM) including an Arduino Uno, L298N/L9110S motor drivers, and 360-degree servos. Conducted a Proof of Concept (POC) using HC-SR04 ultrasonic sensors to trigger servo rotation.

Challenge: Key mechanical components (chassis and motor drivers) experienced shipping delays from Taobao.

Tackle: Pivoted to software-focused development, writing the core sensor-logic scripts in C++ ahead of hardware arrival to simulate movement patterns

Jan 9 - Jan 17, 2026: Problem Definition & Research

Process: Conducted a socio-technical analysis of waste management in Hong Kong. Identified that 87 littering cases occur daily and that current manual cleaning is inefficient and intrusive during daytime hours.

Challenge: Existing robotic solutions are often too small or require frequent recharging, making them unsuitable for industrial street cleaning.

Tackle: Proposed a stationary "hub-and-spoke" model where large-scale units conduct thorough cleaning during off-peak hours (midnight) and return to designated centers for recharging.

Google Sites

Report abuse