PerchScale - Low-cost Modular IoT for Avian Conservation

The Team

The members of our dedicated team form part of the University of Cape Town's Radar Remote Sensing Group (RRSG) and FitzPatrick Institute of Ornithology. Mishay (left), Tristyn (centre left), and Ryan (right) are all passionately working on PerchScale as a part-time project while completing their Master's at the university. Dr. Stephen Paine (centre right) is head of the RRSG and has been the senior supporting member of the team. Mr Ben Murphy, a PhD candidate and passionate ornithologist from the FitzPatrick Institute of Ornithology has been a key advisor and partner on this project.

Overview

The PerchScale project aims to aid avian research efforts by developing a low-cost, modular IoT platform that can be used to gather crucial data on various bird species in remote environments. Initially aimed at autonomously weighing Southern Yellow-billed Hornbills in the harsh conditions of the Kalahari Desert, this collaboration with the FitzPatrick Institute of Ornithology at the University of Cape Town has since evolved into a more versatile platform. Now, PerchScale is set to be an adaptable platform, tailored to accommodate different bird species with unique perch designs that makes use of universal embedded electronics hardware.

Objectives

Comprehensive and accurate weight data is essential for understanding how adult avian body conditions fluctuate in response to environmental stressors. With climate change exerting increasing pressures on bird populations, monitoring weight change is vital for assessing reproductive success and survival rates. As temperatures rise, many species experience significant weight fluctuations, which is closely linked to their ability to forage and thrive. The Southern Yellow-billed Hornbill, for instance, is projected to face extinction within the next 40 years if current trends continue. By tracking weight fluctuations and their correlations with foraging success and reproductive health, researchers can gain valuable insights into the challenges these birds face. This insight can then be used to create potential mitigation strategies for the conservation of these birds and their habitats.

One of the key advancements offered by PerchScale is its ability to automatically extract weight data of wild birds as they land on the scale. Traditional scales often struggle with the rapid movements and light weights of birds, leading to inaccuracies in collected data or the need for manual intervention, which can lead to stress and influence feeding patterns. PerchScale's design accounts for these challenges by using real-time data processing to accurately capture weight measurements, even as the birds land on, move around, and take off from the scale. Due to this, the risk of data errors caused by bird activity is minimised, ensuring that researchers receive precise and reliable data. Moreover, the system's ability to operate autonomously allows for continuous monitoring without the need for human presence, minimising the impact on the birds' natural behaviours and providing a more efficient, user-friendly, and non-invasive solution.

Currently, PerchScale is focused on two primary species: the Southern Yellow-billed Hornbill in the Kalahari Desert and the Red-winged Starling in urban environments. The Red-winged Starling project aims to investigate how urban food sources affect components of fitness — such as nesting growth, reproductive success, and adult weight fluctuations. However, the technology and methodology developed through PerchScale is designed to be adaptable and applicable to a wide range of bird species in various environments.

The objectives of the PerchScale project are:

1. Non-invasive Monitoring: To enable continuous monitoring of weight and dietary habits without disrupting the natural behaviour of the birds, leading to more accurate and reliable data collection.
2. Improved Data Accuracy: To ensure that the system and therefore the datasets collected are properly calibrated to ensure the reliability of the data.
3. Resilience in Harsh Conditions: To create a durable system capable of functioning effectively in extreme environments like the Kalahari Desert, ensuring long-term data collection with minimal maintenance.
4. Affordability and Accessibility: Recognising the financial constraints faced by many academic institutions, particularly in developing countries, our system is designed to be cost-effective. This affordability is crucial for community engagement and large-scale deployment, allowing for the gathering of meaningful data across various species.
5. Seamless Integration: The technology is engineered to fit within existing habitat infrastructures with minimal disruption, promoting easy adoption across different species while remaining compatible with current research practices.

Ultimately, the PerchScale project aims to provide the ornithology research community with vital data on avian weight and health across multiple species. This information is critical for developing effective conservation strategies, allowing researchers to make informed decisions that could reverse the decline of species. By equipping researchers with an affordable, accurate, and user-friendly monitoring system, PerchScale aspires to make a significant impact on avian conservation efforts.

Methodology

PerchScale began as a project for a 4th-year engineering design course at the University of Cape Town, run by Dr. Stephen Paine. The initial goal was to assist Ben Murphy, an ornithologist at the FitzPatrick Institute, with his research on the Southern Yellow-billed Hornbills. The project has since evolved into a broader collaboration with ornithologists at the FitzPatrick Institute, focusing on a multi-species design.

Due to the strict time and resource constraints of the design course, much was left unfinished at the end of 2023. Despite this, the course project provided a proof-of-concept (PoC) prototype which could be attached to the existing Southern Yellow-billed Hornbill nest boxes used by Ben Murphy for his research. This prototype, shown in Figure 1, could be used to obtain weight estimates of birds sitting on the perch.

Following the course, a strong interest emerged to continue the project. Collaborating with the ornithologists, we re-evaluated the requirements for creating a Minimum Viable Product (MVP). Engaging with Ben Murphy and the FitzPatrick Institute revealed a broader demand for this system, prompting PerchScale's evolution into a multi-solution platform for various bird species.

To achieve this new goal, the project’s Research and Development (R&D) has been divided into six work packages, each with specific deliverables outlined in Table 1.

Table 1: Project Deliverables

It is essential to note that all work packages must be revisited for each bird species; this is due to their unique perch design requirements which depend on their environment and perching habits. Ongoing R&D will always be necessary for perch designs the goal for the electronics platform is to create a fully deployable, modular system. This will allow for easy customisation of the sensor suite at the system deployment level (WP 6) for each version, eliminating the need for continued electronic R&D. Once the electronics are finalised, only a new perch and housing will need to be designed for each species.

Currently, the focus is on developing a version of PerchScale tailored for the Starlings, which are being studied at the University of Cape Town’s Upper Campus. This accessible location will facilitate ongoing testing of the electronic hardware. Success with the Starlings will yield the electronic hardware necessary for the Southern Yellow-billed Hornbill version of the system, with a prototype expected to be deployed and tested in late 2024.

The subsequent sections of this article will delve into perch designs specifically crafted for the Starlings. However, the electronic hardware prototypes developed will be applicable to all future versions of PerchScale, regardless of species.

WP 1 Engaging with Stakeholders

Testing the PoC revealed several flaws: the bar perch introduced a bending moment, resulting in inaccurate readings; the perch height allowed birds to eat the bait without stepping on the scale; and the system depended on a connected laptop for power, data storage, and control, hindering autonomous operation. Discussions with the ornithologists revealed that remote control (not a control panel on the system itself) would be advantageous as it would mean less human interference with the birds.

WP 2 Prototype Design

Perch and Housing

The redesign of the PoC prototype led to the T-perch, shown in Figure 2. This design is advantageous as it positions the bird's weight directly over the load cell which decreases the rotational moment, and its increased height prevents birds from accessing bait without using the scale. To support the new perch, the electronics housing was redesigned and is shown in Figure 3. However, testing with the Starlings revealed stability issues, with T-perch v0 tipping over due to unbalanced loads and lacking real-time weight visibility, crucial for assessing system function.

Revisions resulted in T-perch v1, depicted in Figures 4 and 5, where the load cell is centrally positioned within the electronics housing, and an OLED display screen was added to show real-time readings, similar to a regular kitchen scale.

Electronic Hardware

With the perch and housing redesign in place, focus shifted to the electronic hardware. An SD card was incorporated for temporary data storage, and a Real-Time Clock (RTC) module resolved timestamping issues. Various remote control options were evaluated, including a local web server, mobile app, and custom-built controller. After further discussions with Ben Murphy, the custom controller was selected for its simplicity and functionality for fieldwork. Designed with a 4-button D-pad configuration, the controller facilitates menu navigation and displays the scale control menu on an OLED screen, as illustrated in Figure 6.

WP 3 Software/Firmware Development

The firmware and software components are critical for the PerchScale system’s functionality. Key functionalities include accurately reading values from the load cell and converting them to weights, displaying processes and real-time readings on the OLED, and saving timestamped weight readings to the SD card. The controller must display the control menu and facilitate using the 4-button D-pad, employing Bluetooth Low Energy (BLE) for scale instructions. Additionally, both the scale and controller will communicate the State of Charge (SoC) of their power systems to users.

Product

A significant aspect of prototype development for PerchScale involves field testing. A good field-testing kit is an important asset that contributes to the success of both testing and deploying the system. For this reason, we are excited to use the RS Student Project Fund to purchase essential equipment for fieldwork, including calibration weights, screwdrivers, and a multimeter.

Next Steps

Building on the progress made so far, the next steps involve advancing WP 2, 3, and 4 to develop a fully functional prototype for the Red-winged Starlings. Key actions include:

1. Integrating the weighing mechanism with BLE.
2. Completing the design of the controller's electronic hardware.
3. Conducting power budgeting and design for both the controller and scale.
4. Finalising the housing design for the controller and scale.
5. Ensuring the firmware and software function as desired for both devices.
6. Designing a custom printed circuit board for the controller and scale.

Once WP 4 is completed, we will engage with stakeholders to facilitate prompt progress on WP 5 and initiate WP 6 for the Starlings.

Simultaneously, work on WP 2 and WP 3 for the Southern Yellow-billed Hornbills will commence. This phase will involve incorporating a camera and temperature sensors into the electronic hardware, along with the necessary firmware and software adjustments. Additionally, we will redesign the housing to accommodate the existing Hornbill nest boxes, aiming for completion in time for the first pilot deployment (WP 4) in November 2024.

Achieving these objectives will pave the way for advancing the Hornbill PerchScale to WP 6. Ultimately, we aspire to collaborate with other ornithologists on perch designs for additional bird species once the Starling PerchScale has been successfully deployed.

See the full list of Winners of this year's Student Project Fund - South Africa!