Skip to the content
4 layers125 × 94.6 mmDespro Gacor, the printed circuit board in 3D

Autonomous Evacuation Rover (AER) is an integrated mobile robotics platform developed for autonomous exploration, mapping, hazard detection, and evacuation assistance in unknown or hazardous indoor environments.

Size
125 × 94.6 mm
Copper layers
4
Thickness
1.6 mm
Parts
13
Nets
63
Pads
116
Tracks, vias
159, 16
Routed
Yes
KiCad
KiCad 10.0 (file 20260206)
Created
1 Sep 2026
Updated
3 Nov 2026

Overview

Despro Gacor is an open-source printed circuit board by BryanFarras, from the BryanFarras/Desain-Proyek-2 repository on GitHub. It is a 4-layer board of 125 × 94.6 mm with 13 parts on 63 nets, routed with 159 tracks and 16 vias.

It is built around the ESP32-DEVKITC-32E, ROB-14450 and MC001002, with the XT60-M among its other parts. Its parts include 7 connectors, 2 capacitors and 2 ICs. It belongs with the robots and motor control boards and the ESP32 boards of this library.

Main components

4

README

README.md

Autonomous Evacuation Rover (AER)

Autonomous Evacuation Rover (AER) is an integrated mobile robotics platform developed for autonomous exploration, mapping, hazard detection, and evacuation assistance in unknown or hazardous indoor environments. The system combines ROS 2-based autonomous navigation (SLAM and Nav2), real-time deep learning obstacle detection (YOLOv8), an embedded dual-motor controller (ESP32 and TB6612FNG), and a real-time web monitoring dashboard designed for both onboard touchscreen kiosks and remote operator telemetry.


Table of Contents


System Architecture

The robot utilizes a distributed computing model comprising a primary Single Board Computer (Raspberry Pi 4B) and an auxiliary Microcontroller Unit (ESP32 DOIT DevKit V1), linked over high-speed USB-UART serial communication:

text
+-----------------------------------------------------------------------+
|                    Raspberry Pi 4B (Host Computer)                    |
|                                                                       |
|  +---------------------------+       +-----------------------------+  |
|  |     ROS 2 Jazzy Stack     |       |   Computer Vision Pipeline  |  |
|  |  - SLAM Toolbox (Mapping) |       |  - YOLOv8 Obstacle Model    |  |
|  |  - Beluga AMCL (Localize) |       |  - Picamera2 / OpenCV       |  |
|  |  - Nav2 (Path Planning)   |       |  - Real-time Bounding Boxes |  |
|  +-------------+-------------+       +--------------+--------------+  |
|                |                                    |                 |
|                +------------------+-----------------+                 |
|                                   |                                   |
|                      +------------v------------+                      |
|                      |  Flask Monitoring Server|                      |
|                      |  - Telemetry Dashboard  |                      |
|                      |  - Live MJPEG Stream    |                      |
|                      |  - Mission Mode Control |                      |
|                      |  - Emergency Stop / HITL|                      |
|                      +------------+------------+                      |
|                                   |                                   |
|                         [Serial Bridge: 115200]                       |
+-----------------------------------|-----------------------------------+
                                    | USB-UART
+-----------------------------------v-----------------------------------+

Key Features

  • Autonomous Navigation and SLAM: Full ROS 2 navigation integration utilizing slam_toolbox for 2D LiDAR mapping, beluga_amcl for probabilistic particle-filter localization, and nav2 for global/local path planning and dynamic costmaps.
  • Deep Learning Obstacle Detection: Onboard real-time inference using YOLOv8 Nano models tailored for low-power edge compute, supporting Picamera2 on Raspberry Pi OS and OpenCV fallback on desktop workstations.
  • Dedicated Low-Level Motor Control: ESP32 firmware running high-frequency 20 kHz PWM for quiet and responsive motor actuation, eliminating timing jitter and isolating motor power noise from the primary processor.
  • Comprehensive Web Monitoring Interface: Flask-based web dashboard offering live teleoperation, real-time telemetry gauges (odometry, yaw, speed, battery, CPU/RAM stats), MJPEG video feed, dynamic whitelisted system diagnostics, and hardware emergency stop triggers.
  • Touchscreen Kiosk Capability: Optimized layout for 3.2-inch to 7-inch Raspberry Pi LCDs with support for full-screen Chromium kiosk autostart and background systemd service operation.
  • Custom Hardware Infrastructure: Dedicated custom PCB designs (Gerber files provided) for clean power delivery, level shifting, motor driver routing, and sensor connectivity.

Repository Structure

text
.
|-- Computer Vision/              # AI training pipelines and camera inference
|   |-- cv_ws/                    # Computer vision workspace resources
|   |-- requirements.txt          # CV specific Python dependencies
|   |-- run_picam_rover.py        # Real-time YOLOv8 camera inference runner
|   |-- standalone_cv.py          # Standalone video test script
|   |-- train_rover.py            # Model training and export (ONNX, TFLite, NCNN)
|   `-- README_TRAINING_PICAM.md  # Detailed training and deployment documentation
|-- Dataset CV/                   # Training and validation datasets for obstacle detection
|-- Docs/                         # Project documentation and setup guides
|   `-- LANGKAH_PEMBUATAN_MODEL_CV.md # Indonesian guide for CV model creation
|-- Monitoring System/            # Legacy telemetry and monitoring code
|-- New Monitoring System/        # Production Flask monitoring and control dashboard
|   |-- aer-dashboard.service     # Systemd unit file for autostart
|   |-- camera_stream.py          # Multi-backend camera streamer (Picamera2/OpenCV)
|   |-- main.py                   # Primary Flask application server and state machine
|   |-- requirements.txt          # Dashboard Python dependencies
|   |-- ros2_node.py              # ROS 2 rclpy node bridge for telemetry
|   |-- serial_bridge.py          # UART communication driver for ESP32
|   |-- setup.sh                  # Dashboard quick-setup script
|   |-- static/                   # CSS stylesheets, JavaScript files, and assets
|   |-- templates/                # Jinja2 HTML dashboard templates
|   `-- README.md                 # Detailed dashboard documentation
|-- PCB/                          # Hardware design files and manufacturing archives

Hardware Specifications

ComponentModel / SpecificationPurpose
Single Board ComputerRaspberry Pi 4 Model B (4GB / 8GB)High-level compute: ROS 2, SLAM, CV, Web Server
MicrocontrollerESP32 DOIT DevKit V1 (30-pin)Low-level motor control, PWM generation, safety brake
Motor DriverToshiba TB6612FNG Dual H-BridgeDrives left and right DC drive motors
Drive MotorsDC Gear Motors with EncodersDifferential drive locomotive base
Range Sensor2D LiDAR (LDLiDAR STL-27L / LD19)Environment 360-degree range scanning for SLAM
Vision SensorRaspberry Pi Camera Module 2/3 or USB CamVideo stream capture and YOLOv8 object detection
Display3.2-inch to 7-inch LCD DisplayOnboard dashboard for local telemetry and status
Power SupplyLi-Ion 3S / 2S Pack + Buck RegulatorsIndependent 5V logic and 7.4V-12V motor supply rails
Custom PCB2-layer FR4 Custom BoardPower distribution, MCU mounting, and header breakouts

Software Stack and Prerequisites

  • Operating System: Ubuntu 24.04 LTS (Noble Numbat) recommended for ROS 2 Jazzy; Raspberry Pi OS 64-bit (Debian Bookworm) supported for the physical robot deployment.
  • Robotics Middleware: ROS 2 Jazzy Jalisco (or ROS 2 Humble Hawksbill on Ubuntu 22.04).
  • Python Runtime: Python 3.10 or higher.
  • Embedded Toolchain: PlatformIO Core (pio) or PlatformIO IDE in VS Code.
  • Core Python Packages:
    • flask >= 3.0.0
    • psutil >= 5.9.0
    • pyserial >= 3.5
    • opencv-python >= 4.8.0
    • numpy >= 1.24.0
    • ultralytics >= 8.0.0
    • torch, torchvision
    • transforms3d >= 0.4.0
    • websockets >= 12.0

Installation and Environment Setup

Automated Installation

Linux and Raspberry Pi OS (One-Line Execution)

For standard application dependencies (Flask, PySerial, OpenCV, YOLO, system libraries):

bash
chmod +x install_all_dependencies.sh && ./install_all_dependencies.sh

For a complete ROS 2 Jazzy installation with Nav2, Gazebo, and simulation dependencies on Ubuntu 24.04 LTS:

bash
sudo bash install_ros_dependencies.sh
Windows (PowerShell)

Open PowerShell as Administrator in the root repository directory:

powershell
Set-ExecutionPolicy -Scope Process -ExecutionPolicy Bypass
.\install_all_dependencies.ps1

Manual Installation

1. System Packages (Linux)
bash
sudo apt-get update && sudo apt-get install -y \
    python3 python3-pip python3-venv python3-dev \
    build-essential libgl1-mesa-glx libglib2.0-0 \
    libgstreamer1.0-0 v4l-utils

If deploying on Raspberry Pi with the official camera:

bash
sudo apt-get install -y python3-picamera2
2. Python Virtual Environment

Create a virtual environment with --system-site-packages so ROS 2 (rclpy) and picamera2 system bindings remain accessible:

bash
python3 -m venv --system-site-packages .venv
source .venv/bin/activate
pip install --upgrade pip
pip install -r requirements.txt
3. Serial Port Permissions (Linux)

Grant your user access to USB serial devices (/dev/ttyUSB0 or /dev/ttyACM0):

bash
sudo usermod -aG dialout $USER

Log out and back in for this change to take effect.


Subsystem Configuration and Execution

1. Web Monitoring System

The monitoring system runs a Flask server that bridges ROS 2 telemetry, camera video, and direct ESP32 serial communication.

bash
# Activate environment
source .venv/bin/activate

# Launch the monitoring server
cd "New Monitoring System"
python main.py

The interface will be accessible in your browser at: http://localhost:5000 or http://<ROBOT_IP>:5000

Running as a Background System Service (Linux)

To enable the dashboard to start automatically on system boot:

bash
sudo cp "New Monitoring System/aer-dashboard.service" /etc/systemd/system/
sudo systemctl daemon-reload
sudo systemctl enable aer-dashboard.service
sudo systemctl start aer-dashboard.service
Kiosk Mode Configuration (Raspberry Pi Display)

To launch the dashboard full-screen in Chromium on desktop login, create an autostart entry:

bash
mkdir -p ~/.config/autostart
cat << 'EOF' > ~/.config/autostart/aer-dashboard.desktop
[Desktop Entry]
Type=Application
Name=AER Dashboard
Exec=chromium-browser --kiosk --app=http://localhost:5000 --start-fullscreen
EOF

2. Computer Vision and Obstacle Detection

Real-Time Camera Inference

Run the detection pipeline using either a connected Raspberry Pi Camera or standard USB/OpenCV webcam:

bash
cd "Computer Vision"

# Using OpenCV / USB Webcam:
python run_picam_rover.py --cam opencv --cam-id 0

# Using Raspberry Pi Camera Module:
python run_picam_rover.py --cam picam --conf 0.45
Model Training and Export

To train a new YOLOv8 obstacle detection model on your custom dataset:

bash
cd "Computer Vision"
python train_rover.py --epochs 50 --imgsz 416 --batch 16 --model yolov8n.pt

After training, the best model weights will be saved to runs/rover_train/rover_obstacle_model/weights/best.pt and automatically exported to ONNX and TFLite formats for optimized execution on low-power devices.


3. ROS 2 Simulation, SLAM, and Navigation

Building the Simulation Workspace
bash
source /opt/ros/jazzy/setup.bash
cd Rover_Sim
colcon build --symlink-install
source install/setup.bash
Launching Gazebo Simulation

Launch the AREbot robot model within the Gazebo simulation environment:

bash
ros2 launch AREbot_description gazebo.launch.py
Running 2D SLAM Mapping

In a separate terminal:

bash
source Rover_Sim/install/setup.bash
ros2 launch AREbot_description online_async_launch.py
Running Beluga AMCL Localization

To localize the robot within a pre-mapped environment:

bash
source Rover_Sim/install/setup.bash
ros2 launch AREbot_description beluga_localizaton_amcl.launch.py
Launching Nav2 Autonomous Navigation
bash
source Rover_Sim/install/setup.bash
ros2 launch AREbot_description nav2_script.launch.py

4. ESP32 Motor Controller Firmware

The firmware in RoverESP handles motor actuation and receives direction commands from the Raspberry Pi over UART.

Building and Uploading with PlatformIO
  1. Connect the ESP32 to your workstation via USB.
  2. Navigate to the RoverESP directory:
    bash
    cd RoverESP
  3. Compile the firmware:
    bash
    pio run
  4. Upload the binary to the ESP32:
    bash
    pio run --target upload
  5. Open the serial monitor to verify operation:
    bash
    pio device monitor --baud 115200

Communication and Control Protocols

ESP32 UART Serial Protocol

The ESP32 listens on UART (115200 baud, 8N1) for single-character ASCII control commands:

Command CharacterActionDescription
wForwardBoth motors rotate forward at currentSpeed
sBackwardBoth motors rotate backward at currentSpeed
aTurn LeftSteer left (Right motor forward, Left motor reduced)
dTurn RightSteer right (Left motor forward, Right motor reduced)
qSpin LeftCounter-clockwise in-place rotation
eSpin RightClockwise in-place rotation
(Space)Coast StopDisables motor PWM, allowing free rolling
bShort BrakeActuates active brake on both H-bridges
+Speed IncrementIncreases default PWM speed by +20 (Max 255)
-Speed DecrementDecreases default PWM speed by -20 (Min 0)
tAutomated DemoRuns pre-programmed motion self-test
hHelp MenuPrints serial command help interface

Web Monitoring Endpoints

EndpointHTTP MethodFunction
/GETServes the primary monitoring web dashboard
/video_feedGETStreams live MJPEG video with overlay graphics
/statusGETReturns JSON payload of real-time telemetry, sensors, and status
/exec_commandPOSTExecutes pre-whitelisted system diagnostic commands safely
/api/modePOSTChanges the operational mode (STANDBY, PATROL, EVACUATION, MANUAL)
/api/controlPOSTDispatches direct motor commands to the serial bridge

ROS 2 Topic Architecture

Topic NameMessage TypeDescription
/cmd_velgeometry_msgs/msg/TwistVelocity command input for differential drive
/odomnav_msgs/msg/OdometryOdometry state published by controllers
/scansensor_msgs/msg/LaserScan2D LiDAR range measurements
/mapnav_msgs/msg/OccupancyGrid2D SLAM generated grid map
/amcl_posegeometry_msgs/msg/PoseWithCovarianceStampedEstimated robot pose from localization
/joint_statessensor_msgs/msg/JointStateWheel joint angular position and velocity

Hardware Design and PCB

Custom circuit board designs for the Autonomous Evacuation Rover are stored in the PCB directory:

  • PCB/Gerber New/: Latest production-ready Gerber archives including drill files, solder masks, copper traces, and silkscreen layers.
  • PCB/Rover Gerber/: Baseline chassis interface board supporting ESP32 socket mounting, TB6612FNG daughterboard pinouts, logic-level regulation, and direct screw-terminal motor connections.

To manufacture these boards, submit the respective Gerber directory as a ZIP archive to standard PCB fabrication services.


Troubleshooting

Serial Port Access Issues on Linux

If the web monitoring system reports serial connection failure:

  1. Verify device enumeration: ls -l /dev/ttyUSB* /dev/ttyACM*
  2. Add your user account to the serial group: sudo usermod -aG dialout $USER
  3. Reload permissions or reboot the system.

Missing OpenGL / Display Libraries (libGL.so.1)

When running OpenCV on headless Linux systems or minimal server installations:

bash
sudo apt-get install -y libgl1-mesa-glx libglib2.0-0

Camera Device Busy or Not Found

  1. Ensure the camera ribbon cable is firmly seated and enabled in raspi-config (Interface Options -> Camera).
  2. Check available video capture nodes: v4l2-ctl --list-devices
  3. Test using the standalone utility: python "Computer Vision/standalone_cv.py"

ROS 2 Environment Not Sourced

If ROS commands or packages are not recognized:

bash
source /opt/ros/jazzy/setup.bash
source Rover_Sim/install/setup.bash

Ensure these lines are added to your ~/.bashrc file for automatic sourcing in new terminal sessions.


License and Project Information

This project is developed as part of Desain Proyek 2 (Despro). Repository: BryanFarras/Desain-Proyek-2

All code and design documentation are maintained for educational and research purposes in autonomous mobile robotics and disaster response applications.

Parts

13
QtyValueFootprintReferences
2470uCapacitor_THT:CP_Radial_D10.0mm_P5.00mmC1, C2
2Conn_01x04_SocketConnector_PinSocket_2.54mm:PinSocket_1x04_P2.54mm_VerticalJ3, J9
2Conn_01x06_SocketConnector_PinSocket_2.54mm:PinSocket_1x06_P2.54mm_VerticalJ5, J6
1BuzzerBuzzer_Beeper:Buzzer_12x9.5RM7.6BZ1
1Conn_01x08_SocketConnector_PinSocket_2.54mm:PinSocket_1x08_P2.54mm_VerticalJ4
1Conn_01x20_SocketConnector_PinSocket_2.54mm:PinSocket_1x20_P2.54mm_VerticalJ1
1ESP32-DEVKITC-32EESP32-DEVKITC-32E:MODULE_ESP32-DEVKITC-32EU1
1MC00100MC00100:CONV_MC00100PS1
1ROB-14450ROB-14450:MODULE_ROB-14450U2
1XT60-MXT60-M:AMASS_XT60-MJ2

Similar boards

  • 2 layers100 × 70 mmOTO SERA POWER MANEGEMENT BOARD, the printed circuit board in 3D

    OTO SERA POWER MANEGEMENT BOARD

    mbuyukeskil/kicad-hardware-projects

    KiCad PCB design projects, schematics, layouts, 3D views and manufacturing files.

    15 parts
  • 2 layersPcb-2, the printed circuit board in 3D

    Pcb-2

    Khibita/Pendulo_Invertido_TC

    Práctica 4 y 6 de teoría de control

    9 parts
  • 2 layers120 × 100 mmRoot, the printed circuit board in 3D

    Root

    GonzaCerv/noah-hardware

    This repo contains all the hardware description for the Noah project. Here lies 3D models, electronics PCB and Firmware software of the robot.

    127 parts28MIT
  • 2 layers80 × 65 mmPami n20 V2, the printed circuit board in 3D

    Pami n20 V2

    Rami-Troudi/Electrique-Eurobot-IEEE-INSAT-2026

    Repo for my electrical work for the Eurobot competition

    35 parts1
  • 2 layers61 × 74 mmProbador Bobinas PCB, the printed circuit board in 3D

    Probador Bobinas PCB

    ramiroedson72-cloud/probador-bobinas-cop

    Generador de disparos PWM para probar bobinas de encendido (COP). Un potenciómetro controla la frecuencia (10–200 Hz ≈ 300–6000 RPM en motor de 4 cilindros) y el firmware mantiene un dwell fijo de 3 ms recalculando el…

    16 parts1
  • 2 layers131 × 90 mmStage chrono, the printed circuit board in 3D

    Stage chrono

    becem0/esp32-giant-chronometer

    ?? Giant 6-digit 7-segment race chronometer powered by ESP32, 24V COB LED strips, custom KiCad 10 PCB, and SIM800L GSM telemetry.

    38 partsMIT
  • 2 layers240 × 95 mmMotorDriver Board, the printed circuit board in 3D

    MotorDriver Board

    domenicojbruno/IQP-Resonant-Space

    An interactive "resonant wall" for a hallway. VL53L0X time-of-flight sensors detect when people cross a threshold, and a honeycomb of 13 hexagonal panels, each moved by a linear actuator, pulls back from and pushes out…

    30 parts
  • 2 layers66 × 32 mmEnergy, the printed circuit board in 3D

    Energy

    MSMRo/Fluorimetro_lowcost

    La PCB Energy se usa para la alimentación del montaje, con una línea de 5 V para el Arduino Mega y una línea de 12 V para la iluminación.

    13 parts2

Questions

How big is the Despro Gacor PCB?

It measures 125 × 94.6 mm and has 4 copper layers, 1.6 mm thick.

What parts does the Despro Gacor use?

13 parts, built around the ESP32-DEVKITC-32E, ROB-14450, MC001002 and XT60-M. The full list of its parts is on this page, with links to each part's symbol and footprint where PCBKit's library has them.

Where can I download the Despro Gacor KiCad files?

From the BryanFarras/Desain-Proyek-2 repository on GitHub, at the commit this page was made from: the board is PCB/Gerber New/Despro Gacor.kicad_pcb and its schematic PCB/Gerber New/Despro Gacor.kicad_sch. PCBKit opens the board in the browser, with nothing to install.

Can I use the Despro Gacor design in my own project?

Its repository states no licence, so its author keeps all rights: ask BryanFarras before reusing the design.