PilotClean: The Ultimate FPGA Dual-Mode Cleaning Robot

Click on the picture to view our full presentation PDF

Overview

PilotClean is a dual-mode cleaning robot designed and developed by Team 23 in Hardware Design and Lab, a class in NTHU. It leverages FPGA technology to combine the flexibility of manual control with the efficiency of automatic cleaning, addressing limitations in existing robot vacuums. The system integrates hardware and software innovations, including remote control functionality, ultrasonic obstacle detection, and customizable speed settings.

Features

1. App Control:

   • Compatible with Arduino Bluetooth Controller for controlling movement, power, suction, mode selection, and speed levels (LOW, MID, HIGH).

2. Dual Cleaning Modes:

   • Remote-Control Mode: Operated via a smartphone app for precise spot cleaning.
   • Random Mode: Uses Linear Feedback Shift Register (LFSR) for autonomous, random path generation.

3. Obstacle Detection:

   • Ultrasonic sensors (HC-SR04) detect and avoid obstacles to prevent collisions and ensure smooth operation.

4. Dust Collection:

   • Powered by an 8x8 12V 0.6A square fan for efficient suction. The design eliminates the middle filter layer to enhance airflow while protecting internal components.

5. Trash Disposal:

   • Features a return-to-home system for depositing trash at the base station, controlled via remote.

---

System Specifications

• Power Supply: Three 18650 lithium-ion batteries connected in series, delivering a maximum voltage of 12V.
• Processing Unit: FPGA board with a complex Dupont wiring setup.
• Communication: Bluetooth modules (HC-05/HC-06) using the UART protocol.
• Motor Control: Speed and direction adjustments using ultrasonic sensors.

---

Challenges Overcome

1. Dust Management:
   • Removed the filter paper layer to improve suction while keeping an outer masking cover to block large debris.
2. Signal Stability:
   • Implemented OnePulse technology to stabilize UART communication and prevent repeated signals.
3. Random Walk Generation:
   • Redesigned the LFSR logic to include all bits and improved the initial seed for better randomness.

---

Experimental Results

• Suction Performance: Removing the filter paper layer increased airflow significantly.
• Path Randomness: Enhanced LFSR logic resulted in more diverse cleaning paths.
• Power Limitation: Tests revealed that 12V power is slightly underpowered, indicating potential for future optimization.

---

Future Improvements

• Upgrade power system for higher performance.
• Implement fully autonomous return-to-home functionality.
• Enhance app features for greater user convenience.
• Add more advanced sensors for improved obstacle detection and navigation.

---

Usage

1. Clone the repository:

   git clone https://github.com/vaclisinc/PilotClean.git

2. Navigate to the project directory:

   cd PilotClean

3. Feel free to use the source code:
   • The source code files are located in the source_code folder.
   • The constraints file and pin design picture are located in the xdc folder.

---

Team Members

• Song-Ze, Yu (@vaclisinc): Verilog coding, prototype design, visual design, and report writing.
• Yun-Zhong, Lai (@edwinlai2058): UART sub-module coding, hardware assembly, block diagrams, and report writing.