Pyroscope

Autonomous wildfire fuel-load mapping rover - GPS-guided, lidar-navigated, thermal-camera-equipped.

Full resolution: Poster.pdf

---

This project is grounded in field conditions near Ellensburg, WA, about 2 hours from Seattle.

One really good place to look at to assess wildfire risk are weather stations. However, weather stations are based mostly on dead-fuels, and there is little signal from actual live-fuel conditions. Pyroscope is designed to fill that gap by providing a mobile platform for collecting live-fuel data by measuring the temperature of the foilage on the ground vs the surrounding air to have a proxy for soil moisture levels.

Land managers in shrubland ecosystems need reliable surface fuel-load data to plan prescribed burns and reduce wildfire risk, but manual ground sampling is slow, expensive, and hard to scale. A major wildfire in 2012 burned much of this area, highlighting the need for better field intelligence.

Pyroscope was built to close that gap.

Pyroscope is a compact rover that carries thermal and environmental sensors and streams mission data to a dashboard in near real time.

What This Repository Contains

• ROS navigation and sensor stack: catkin_ws/src/...
• FastAPI backend APIs and mission control: application/backend
• React dashboard frontend: application

Project Structure

pyroscope/
├── assets/
│   ├── Poster.png
│   ├── Poster.pdf
│   └── ...
├── catkin_ws/
│   └── src/
│       ├── pyroscope_navigation/      # Coverage planner, move_base launch, safety nodes
│       ├── pyroscope_sensors/         # Sensor launch/scripts (SHT40, thermal, camera)
│       ├── transbot_bringup/          # Base driver + odometry + TF
│       ├── transbot_ctrl/             # Teleop tools
│       └── transbot_nav/              # Legacy map-based nav package
├── application/
│   ├── backend/                       # FastAPI app, DB models, ROS sensor bridge
│   ├── src/                           # React dashboard source
│   └── package.json
├── SENSOR_INTEGRATION_SETUP.md
└── README.md

System Layout

Typical runtime uses multiple machines:

• Jetson (robot): base bringup, lidar, onboard sensors
• Remote Ubuntu PC (same ROS network): move_base + coverage planner (and optionally backend)
• Frontend machine: React dashboard

Main autonomy path:

1. pyroscope_sensors/launch/jetson_bringup.launch starts base + lidar + sensors.
2. pyroscope_navigation/launch/coverage_mission_nav.launch starts move_base and coverage planner.
3. Coverage planner sends waypoints; move_base performs local obstacle avoidance with DWA.

Prerequisites

• Ubuntu 18.04 with ROS Melodic for the ROS stack
• Python 3.9+ for backend
• Node.js 18+ for frontend
• MySQL 8+ for backend persistence

Install ROS navigation deps on the machine running move_base:

sudo apt install ros-melodic-move-base ros-melodic-dwa-local-planner \
  ros-melodic-navfn ros-melodic-move-base-msgs

One-Time Setup

1) Build catkin workspace

cd ~/pyroscope/catkin_ws
catkin_make
source devel/setup.bash

2) Backend setup

cd ~/pyroscope/application/backend
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
alembic upgrade head

Set environment in application/backend/.env (minimum):

DATABASE_URL=mysql+pymysql://<user>:<password>@localhost:3306/pyroscope_db
JWT_SECRET_KEY=<your-secret>
ROS_MASTER_URI=http://<JETSON_IP>:11311
ROS_IP=<BACKEND_PC_IP>

3) Frontend setup

cd ~/pyroscope/application
npm install

Note: API base URL is currently hardcoded in application/src/services/api.js. Update it for your backend host.

How To Run

A) ROS Navigation + Sensors (no web stack)

Terminal 1 (Jetson):

source ~/pyroscope/catkin_ws/devel/setup.bash
roslaunch pyroscope_sensors jetson_bringup.launch # this is a consolidated launch file that launches all the sensors and brings up the robot sensors, lidar, and bringup

This launches:

• transbot_bringup/bringup.launch
• rplidar_ros/rplidar.launch
• pyroscope_sensors/sensors.launch

Terminal 2 (Remote PC or same ROS machine):

source ~/pyroscope/catkin_ws/devel/setup.bash

# the front end takes care of this, but this could be manually run as well: 
roslaunch pyroscope_navigation coverage_mission_nav.launch \
  area_width:=3.0 area_height:=3.0 row_spacing:=1.0 waypoint_spacing:=1.0 \
  origin_x:=0.0 origin_y:=0.0 dwell_time:=3.0 waypoint_timeout:=60.0

What to expect:

• move_base topics appear (/move_base/status, costmaps, plans)
• /coverage/progress advances from 0/N to N/N
• Robot follows a lawnmower pattern centered around (origin_x, origin_y) in odom
• If waypoints fail repeatedly, planner attempts recovery (clear costmaps) and continues mission

Quick checks:

rostopic echo /coverage/progress
rostopic echo /move_base/status
rosrun tf tf_monitor

Expected TF chain: odom -> base_link -> laser

Emergency cancel current goal:

rostopic pub -1 /move_base/cancel actionlib_msgs/GoalID '{}'

B) Full Stack (ROS + Backend + Frontend)

Terminal 1 (Jetson):

source ~/pyroscope/catkin_ws/devel/setup.bash
roslaunch pyroscope_sensors jetson_bringup.launch

Terminal 2 (Backend host):

cd ~/pyroscope/application/backend
source venv/bin/activate
python run.py

Terminal 3 (Frontend host):

cd ~/pyroscope/application
npm run dev

Open dashboard: http://localhost:5173

What to expect:

• Backend on :8000 with /docs and /health
• If ROS_MASTER_URI is set, ROS sensor bridge auto-starts at backend startup
• /api/sensors/availability returns available: true when required topics are present
• Dashboard shows live sensor data and mission controls

API sanity checks:

curl http://localhost:8000/health
curl http://localhost:8000/api/sensors/availability
curl http://localhost:8000/api/robot/mission/status

Mission API Quickstart

Start mission:

curl -X POST http://localhost:8000/api/robot/mission/start \
  -H "Content-Type: application/json" \
  -d '{"area_width":3.0,"area_height":3.0,"row_spacing":1.0,"waypoint_spacing":1.0,"origin_x":0.0,"origin_y":0.0}'

Check status:

curl http://localhost:8000/api/robot/mission/status

Stop mission:

curl -X POST http://localhost:8000/api/robot/mission/stop

Other Useful Launch Modes

Keyboard teleop:

roslaunch transbot_ctrl transbot_keyboard.launch

Manual waypoint controller:

roslaunch pyroscope_navigation exploration.launch
rosrun pyroscope_navigation test_single_waypoint.py 2.0 1.0

Random exploration:

roslaunch pyroscope_navigation random_exploration.launch

Standalone safety layer:

roslaunch pyroscope_navigation obstacle_avoidance.launch

Key Launch Files

Launch file	Purpose
catkin_ws/src/pyroscope_sensors/launch/jetson_bringup.launch	One-command robot bringup (base + lidar + sensors)
catkin_ws/src/pyroscope_navigation/launch/coverage_mission_nav.launch	Primary autonomous coverage (move_base + DWA)
catkin_ws/src/pyroscope_navigation/launch/exploration.launch	Manual waypoint navigation
catkin_ws/src/pyroscope_navigation/launch/random_exploration.launch	Random walk mobility test mode
catkin_ws/src/pyroscope_navigation/launch/obstacle_avoidance.launch	Lidar obstacle detector + safety stop

Troubleshooting

• move_base not found:
  • Install ROS nav packages listed in Prerequisites.
• Planner waits for action server then aborts:
  • Confirm coverage_mission_nav.launch is running and TF chain is valid.
• Sensors unavailable in backend:
  • Verify ROS_MASTER_URI/ROS_IP, then run:
    • application/backend/scripts/diagnose_sensor_availability.sh
• Odometry missing:
  • Run catkin_ws/src/pyroscope_navigation/scripts/check_topics.sh

Known Caveats

• Use coverage_mission_nav.launch as the primary autonomous mode.
• coverage_mission.launch is legacy and not the preferred demo path.
• Very small mission areas can trigger edge-case waypoint fallback behavior.

Additional Documentation

• catkin_ws/src/pyroscope_navigation/README.md
• catkin_ws/src/pyroscope_sensors/README.md
• application/backend/SETUP.md
• application/backend/ROS_SENSOR_SETUP.md
• SENSOR_INTEGRATION_SETUP.md