Bicycle Mounted Pollution Sensor
Over the summer, a colleague and I developed a pollution sensor intended for integration with bike-sharing programs. The device measures time and position-resolved data of common urban pollutants (CO2, NOx, Particulate Matter) providing high spacial resolution data to cities. We are featured on BU's News Letter
2024 Update: BU News Letter
Schematic of Sensors. The device consists of five Sensors, a GPS modem used for location and time, a MicroSD Card to log data, and a 3200 mAh Lithium Ion Battery pack for autonomy. The entire system is controlled by an Arduino Nano
The device is built around an Arduino Nano, facilitating real-time data collection from five core sensors that measure: particulate matter, CO₂, methane, NO₂, temperature and humidity. To obtain accurate position and time data, the system uses a Neo-6M GPS module.
This enables each data point to be geotagged, providing both spatial and temporal context. As a result, pollution levels can be mapped, making it possible to visualize when and where specific pollutants are present.
Hardware Overview
The software begins by initializing and verifying communication with all sensors. Analog sensors are read directly via the Nano’s ADC, while digital sensors communicate over I2C or Serial interfaces. Once all connections are confirmed, the system enters a timed loop, collecting sensor data every 5 seconds. Each reading is paired with GPS-derived location and time data, then formatted into a structured output. This output is transmitted via Serial to a connected device, where it can be logged, visualized, or further processed for analysis.
Onboard Software Overview
%20(1).png)
Code Flow Chart. To reduce startup transient spikes, it contains a delay for the first 30 seconds. After which, a loop starts up where measurements are sampled every 5 seconds and written to a CSV
Since the project team consisted of just two people (myself & Lorenzo Barale) I was deeply involved in nearly every aspect of the work. This included project planning, conducting preliminary research to identify the most critical pollutants, and evaluating low-cost sensors for performance.
I also designed the sensor enclosure in SOLIDWORKS and created the PCB in KiCad, and of course helped solder and assemble everything together. I also contributed to developing the onboard software and built MATLAB apps to analyze the collected data. Overall, it was a very multifaced project.
My Role
Final Device Interior and Exterior. Custom PCB and Compact Enclosure with mesh for airflow. It is designed to be mounted to a Blue Bike (Bostons Bike Share System)
PCB Designed in KiCad and manufactured by PCBWay.
Features an LED to indicate when a measurement is taken.
Pollution Graphs
Due to the limited 10-week timeframe, we were unable to draw large-scale conclusions. However, we did identify two distinct areas with elevated CH₄ levels, suggesting potential localized sources of methane emissions worth further investigation.
Post data collection software to quickly analyze and map methane levels
BU Metcalf Methane Leak
Roxbury Methane Leak
Conclusion and Legacy
I was very happy with the progress we made in only 10 weeks, and it shows how much can be done when you have a strong focus. Since then, we have passed the project on to Senior Design teams to continue development in 2024 and 2025.
Awards:
Janetos Climate Action Prize (2023)
Senior Design: Best Project in Energy & Environmental (2024)
Student Sustainability Leadership Award (2024)
Climate Action Prize Photo (Original team + Senior Design team of 2024)