1 of 9

Team A1: Fire Away

Karen Abruzzo, Ankita Bhanjois, Arden Diakhate-Palme

2 of 9

Use Case

  • In 2022, California had a total of 6,465 wildfires recorded, spanning over 300K acres
  • Our sensor network will wirelessly be triggered in order to notify the public of the locations of wildfires
    • Sensors will gather data about related CO2 levels, temperature, and air humidity around a node to determine where fires are located
  • Sensor Network can be scaled

​

  • ECE Areas: Software & Hardware Systems

3 of 9

Use Case Requirements

​

  • Low Power Nodes
    • Lifetime of at least 34 hours
  • Fire Detection Accuracy
    • Be able to detect the fires with > 95% accuracy
  • Fire Detection Notification Timing
    • Notify the web application of active fires within real time of sensor readings

1-60 min

1-3 hrs

days

Address fire before it becomes an open fire

4 of 9

Technical Challenges

  • Maintain low power design of wildfire sensor nodes
  • Scheduling sensor node wakeup to collect sensor data and send data to router
  • low-latency data transmission and power-efficient routing across mesh network to border gateway router
  • Mesh network resiliency to link failures
  • Be able to notify the web-application of fire detection without (significant) delay
    • Real time data will be received from the embedded system and will need to be relayed to web app

​

5 of 9

Solution Approach

​

  • Sensors that will be used: Wind speed, temperature, CO2, air humidity and air pressure
    • We will be using 8-10 nodes to simulate our system on smaller scale
  • The tools we will be using is:
    • STM32
    • Various Sensors

​

6 of 9

Solution Approach (cont.)

​

STM32L4xx

LoRa transceiver

BME-688

LiPo

UART

SPI / I2C

3.3V

3.3V

Node Architecture

humidity with ±3% accuracy

barometric pressure with ±1 hPa absolute accuracy

temperature with ±1.0°C accuracy

7 of 9

Sensor

Node

Sensor

Node

Sensor

Node

Sensor

Node

Sensor

Node

Sensor

Node

Sensor

Node

Sensor

Node

Border gateway

RPi

Web Application

8 of 9

Testing, Verification, and Metrics

​

  • Testing Fire Detection
    • Turn on a lighter near one of the sensor nodes to trigger the sensors to start triggering the sensor network
    • Testing how quickly the sensor network begins signalling other nodes
    • Testing how many trials are signalled correctly
  • Testing RTOS/Web Application
    • When triggering the system, does the web application show all the data needed?
    • Latency: how long from when fire is triggered that the web application displays information

​

9 of 9

Tasks & Division of Labor

  • Embedded System and RTOS (Karen & Arden)
    • our system needs to retrieve data from the sensors in real time
    • we will be using the following components (SW & HW):
  • Sensor network and Routing System (Arden & Ankita)
    • our system needs to take in 8-10 node and be able to communicate amongst one another and with the central node
    • this system is solely done with software
  • Web Application (Ankita)
    • our web application is mainly for display purposes and notifications/alerts for the general public