Airflow Source Seeking on Small Quadrotors Using a Single Flow Sensor

Lenworth Thomas, Tjaden Bridges, Sarah Bergbreiter · 2026

Read Airflow Source Seeking on Small Quadrotors Using a Single Flow Sensor by Lenworth Thomas, Tjaden Bridges, Sarah Bergbreiter (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-15607.txt.

Introduction

Lenworth Thomas, Tjaden Bridges, and Sarah Bergbreiter teach small drones to sniff out sources in a 2026 arXiv paper. After disasters, finding the source of pollutants or harmful particles matters fast. Plume tracking is the standard approach to that hunt. Small quadrotors could work near people and in tight spaces. Their size is also the problem. Gas sensors that fit them are often slow and insensitive.

The team takes a different route: follow the airflow itself. Their quadrotor uses a single flow sensor to seek the source. Moving air carries the plume's signature, and flow can be sensed quickly. A single cheap sensor keeps the drone light and simple. For search teams, a nimble source-seeker could shorten the time to find leaks and hazards.

Catalog record

CREATORLenworth Thomas, Tjaden Bridges, Sarah Bergbreiter
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE26 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-15607

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Airflow Source Seeking on Small Quadrotors Using a Single Flow Sensor Lenworth Thomas, Tjaden Bridges, Sarah Bergbreiter Abstract As environmental disasters happen more frequently and severely, seeking the source of pollutants or harmful particulates using plume tracking becomes even more important. Plume tracking on small quadrotors would allow these systems to operate around humans and fly in more confined spaces, but can be challenging due to poor sensitivity and long response times from gas sensors that fit on small quadrotors. In this work, we present an approach to complement chemical plume tracking with airflow source-seeking behavior using a custom flow sensor that can sense both airflow magnitude and direction on small quadrotors ( < < 100 g 100\text{\,}\mathrm{g} ). We use this sensor to implement a modified version of the ‘Cast and Surge’ algorithm that takes advantage of flow direction sensing to find and navigate towards flow sources. A series of characterization experiments verified that the system can detect airflow while in flight and reorient the quadrotor toward the airflow. Several trials with random starting locations and orientations were used to show that our source-seeking algorithm can reliably find a flow source. This work aims to provide a foundation for future platforms that can use flow sensors in concert with other sensors to enable richer plume tracking data collection and source-seeking. I Introduction Understanding the sources and patterns of airflow can provide essential insights into the source and dispersion of gases or airborne particulates [ 21 ] . This can be helpful in identifying the source of gas leaks after natural disasters [ 17 ] , environmental monitoring [ 24 ] , and hazardous emissions from waste products [ 6 ] among other applications. Given the hazard to humans in these applications, plume tracking has a rich history in the robotic space on ground-based vehicles [ 5 ] , underwater vehicles [ 19 ] , and unmanned aerial vehicles (UAVs) [ 25 ] . Plume tracking aims to detect, quantify, and, in some cases, follow the chemical or particle signals of interest to their source, a behavior that we call ‘source-seeking’. UAVs are particularly interesting because they can seek sources in 3D space and cover larger areas than may be feasible with fixed sensors [ 21 ] . Recently, many research efforts have focused on reducing the size of plume tracking and source-seeking UAVs. Smaller UAVs (e.g., < 100\text{\,}\mathrm{g} ) can operate in more spatially constrained environments and are safer to operate near humans. Given the typical lower cost of these small robots, they can also be used in swarms that can cover larger areas with higher sensing resolution. There have been several examples of small quadrotors designed for source-seeking behaviors [ 20 ] [ 25 ] [ 7 ] [ 9 ] . These systems often use commercial gas sensors, custom-engineered flow sensors [ 11 ] , or arrays of gas sensors to find gradients in the chemical signature. A significant challenge in these systems is that gas sensors respond slowly, resulting in long times to find a gradient and track it to its source. Gas sensors can also be significantly affected by rotor flow induced in UAVs [ 14 ] . Fig. 1: (a) A flying DJI Tello quadrotor with the flow sensor mounted on top. (b) Exploded view of the flow sensor and quadrotor mount. Several recent approaches have been designed to overcome these challenges of detection speed and sensitivity for source-seeking behaviors on small quadrotors. Anderson et al. used a bio-hybrid approach by directly interfacing a moth antenna to the quadrotor as a chemical sensor [ 2 ] [ 3 ] . The moth antenna enabled a fast response time but only has a limited set of detectable odors. In another approach, a swarm of small UAVs equipped with commercial gas sensors was used to find an alcohol source over a relatively larger area cluttered with obstacles [ 13 ] . This approach, Sniffy Bug, also took advantage of a novel gas dispersion model implemented on the small quadrotors [ 13 ] . Other approaches use novel algorithms to overcome slow and noisy sensors. For example, researchers have used reinforcement learning on small quadrotors for source-seeking tasks, although the source in this case was a light instead of a gas [ 12 ] . Another approach to overcoming the slow response of gas sensors on small UAVs for source-seeking tasks is to use measurement of the airflow itself to aid detection. This idea of anemotaxis (following a flow source) to complement chemotaxis (following a chemical source gradient) is hypothesized to be a primary mechanism in aiding insects in their exceptional odor-seeking skills [ 4 ] [ 27 ] . There are many methods by which to detect airflow on a UAV as reviewed in [ 1 ] , but many of these methods and sensors are only suitable for larger UAVs. Small UAVs can use pose estimation approaches that use IMU measurements combined with commanded rotor speeds, but these are often susceptible to significant drift over time [ 28 ] . Custom flow sensors can also be designed for small UAVs [ 18 ] [ 10 ] , but obtaining high quality flow data often requires multiple sensors spread over a large area, as seen in the large UAV in [ 26 ] . The primary contribution of this work is the integration of a single flow sensor and a bioinspired source-seeking algorithm on a small palm-sized quadrotor for airflow source-seeking (Fig. 1 ). The flow sensor outputs both the magnitude and direction of airflow to enable rapid detection and response. After characterizing this sensor in flight tests, we demonstrate that we can quickly reorient the quadrotor toward the airflow source. The goal of this ‘flow servoing’ approach is to ultimately aid future on-board (and often directional) gas or other environmental sensors by increasing their sensitivity and response time. These results are combined in demonstrations of a bio-inspired algorithm, Vector Surge, that is designed to find airflow in the environment and follow the airflow upstream to its source. This algorithm builds on previous bio-inspired algorithms like ‘Cast and Surge’ [ 3 ] by incorporating the airflow direction obtained by the flow sensor. II Integrating a Flow Sensor With a Small Quadrotor II-A Flow Sensor Background The design of the flow sensor used in this project (Fig. 1 ) was heavily inspired by the work of Kim et al. [ 18 ] . Airflow on the sensor fin creates a moment that rotates the fin and a magnet placed at the bottom of the fin with respect to a planar spring suspension. A magnetic field sensor is placed underneath the fin to measure the change in magnetic field that occurs from the displacement and rotation of the fin magnet in three axes, providing an estimate of the fin rotation. The magnitude of the rotation provides an estimate of the incident airflow velocity, while the direction of the fin movement provides an estimate of the incident airflow direction. This magnetic field measurement (in milliTesla) can be translated to rotation (in degrees) and airflow velocity (in meters/second) as seen in [ 18 ] , but for this work, we

Read the complete text