Drones in Education: A Comparative Analysis of Learning-Focused Platforms

1. Introduction

Unoccupied Aerial Vehicles (UAVs), commonly referred to as drones, have emerged as powerful tools across various applications, including military operations [1], commercial industries, and agriculture [2,3]. A UAV is an aircraft that operates without a human pilot onboard, instead being remotely controlled or operating autonomously through pre-programmed flight paths and artificial intelligence-based navigation systems. Equipped with sensors, cameras, and communication systems, UAVs can perform a wide range of tasks that were once limited to manned aircraft, offering cost-effective, flexible, and accessible solutions for aerial operations.

UAV technology has rapidly advanced and become accessible for civilian applications. Today, drones are used in industries such as agriculture, logistics, filmmaking [4], environmental monitoring [5], and infrastructure inspection[6,7]. Their ability to capture real-time data, reach inaccessible locations, and perform automated tasks has positioned them as valuable assets across multiple sectors.

In recent years, UAVs have also gained significant attention in education [8], offering new ways to enhance learning experiences through hands-on, technology-driven applications.

Integrating UAVs into educational settings has introduced an innovative approach to teaching subjects such as engineering, computer science, and environmental studies. Drones provide students with experiential learning opportunities, allowing them to engage in real-world applications of concepts like aerodynamics, coding, and data analysis. Moreover, UAVs encourage interdisciplinary learning by bridging technical subjects with creative fields like digital media, geography, and environmental research. As drone technology becomes more advanced and accessible, it presents a unique opportunity to cultivate critical skills such as problem-solving, teamwork, and innovation among students.

Despite the growing adoption of UAVs in education, the landscape of drone-based learning platforms remains diverse, with various options available for different levels of instruction and application areas. Some UAV platforms focus on introductory programming and flight mechanics, while others emphasize autonomous navigation, machine learning integration, or specialized applications like aerial mapping and environmental sensing. Understanding the capabilities, limitations, and effectiveness of different educational UAV platforms is essential for educators and institutions aiming to incorporate drones into their curricula.

This paper analyzes various UAV platforms designed for educational use, examining their features, usability, and learning outcomes. By evaluating the strengths and challenges of different drone-based learning tools, this study offers insights into the best approaches for integrating UAV technology into education, ensuring that students and educators can maximize the benefits of this emerging technology.

2. Drone Platforms for Education

Prior research has indicated the substantial benefits of integrating robotics into the curriculum as early as the 8th grade [9]. Robotics engages the students in a creative and practical viewpoint, allowing them to conceive situations and solutions beyond theoretical knowledge and apply them to real-world applications. UAVs can take these benefits further as they have the novelty of being airborne and represent an advanced culmination of the robotics field with their array of sensors, physical structure, and abilities. Integrating UAV awareness in high school and college education through two primary methods is possible. Using simulation platforms is the first method [10]. Often, the first step involves testing the created code in a sandbox environment. Modern-day simulators offer intuitive features such as no-code mission creation, side-by-side action-reaction results, and step-by-step learning modules that can be easily adapted in a classroom environment.

The second method is using UAV hardware models that have been curated explicitly with student education in focus. These user-friendly UAV platforms integrate safety features such as break-resistant materials, soft propellors, propellor guards, and modular builds, ensuring a secure learning environment. Students can explore real-time data collection, geographical mapping, and environmental monitoring with built-in cameras and sensors. These platforms foster hands-on learning in STEM subjects, allowing students to understand complex concepts through interactive experiences. The current market has multiple such platforms that have been created, ready to cater to classroom needs.

They come at various price points and are equipped with competitive features. As such, choosing the right platform for a classroom can be challenging. The authors acknowledge that selecting an EDU platform can be highly subjective based on program budget, needs, and learning objectives. As such, this study focuses on providing a detailed comparison and fact sheet for the three popular drone-based EDU platforms rather than scoring them on a standard scale. The three platforms are the Tello EDU drone from DJI, the CoDrone EDU platform from Robolink, and the Crazyflie drone from Bitcraze. Figure 1 shows a new model of the Robolink CoDrone EDU platform flying outdoors at a height of ∼25 meters in the presence of 2 m/s winds.

The drones are compared over several factors, such as hardware, capability, physical facts, usage, and support characteristics.

2.1. Platform and Company Peripherals

The three platforms are competitively priced in the US market. They are also supported by a strong and active community. DJI and Robolink also offer additional vehicle types, such as other drone platforms and ground vehicles. Table 1 highlights some high level information for the three platforms.

2.2. Hardware

Table 2 compares the three platforms with regards to their hardware characteristics. The drone platforms must be built of robust materials to allow students to experiment freely without fear of constant damage to hardware. Another factor is having a modular build for these drones, so the entire platform does not need to be replaced if one component is damaged. Replacement parts can be ordered from the manufacturer and swapped with the damaged part, keeping overall costs in check. Modular builds allow specific parts to be easily upgraded, allowing the unit to keep pace with curriculum objectives and research. The DJI Tello falls back in this regard with a unibody build. The propellers and motors are the only two components that can be swapped. The battery is removable on all three platforms. The Robolink CoDrone and the Crazyflie are modular to a larger extent, allowing additional component replacement and upgrades. The Crazyflie has expansion decks that can be added to the drone to improve capabilities. Radio modules and processing units can also be swapped. The CoDrone has a higher replaceability and repairability rating than the Tello but has not featured any expansion decks. Sensor customization is limited, as additional sensors cannot be added. The hardware of the CoDrone, unlike the Crazyflie, is also closed source. However, due to its forward-facing camera, the Tello has a feature that is not standard with the other two platforms. It offers a live 720 pixel resolution flight view where the drone can stream its forward-facing camera view directly to the users’ control device. Figure 2 shows a use-case scenario of the live view from the DJI Tello, where the authors used a rescuer agent’s camera feed to search for drone agents in distress [11].

The Tello has decent camera specifications for an inexpensive drone. Multiple project possibilities exist for using Tello’s camera and live view for applications such as marker recognition and navigation, object detection [13], tracking [14], and target following. Table 3 outlines the Tello forward-facing camera specifications.

It is worth noting that while Crazyflie 2.X uses brushed DC motors, Bitcraze has also released a newer version, Crazyflie 2.1, which uses brushless DC motors. These offer improved efficiency, longer flight times, and higher thrust. The specifications outlined in this study refer to the newer model. While only the CoDrone comes with a controller in the box, the Crazyflie and the Tello have a mobile application for iOS and Android that can control the drone. Specific Bluetooth gamepad controllers can also be used with the Tello and the Crazyflie. The CoDrone platform does not support direct connection with a phone. However, all three platforms can be connected to a laptop computer to control them. In such scenarios, web-based tools such as Blockly are suitable for controlling these drones.

Additionally, third-party tools like MATLAB support drone integration and control. For example, MATLAB has an independent support package for connecting and controlling Tello drones. With multiple features and an informative GUI, it is an ideal way to code Tello drones and create additional applications for them. Figure 3 shows a view from the MATLAB support package to connect and control Tello drones.

None of the three platforms can use GNSS (Global Navigation Satellite System). However, there are some interesting workarounds for this shortcoming.

The Crazyflie has two methods for positioning. One is the Loco positioning system, which consists of anchors and base stations. The anchors act like satellites in a GPS (Global Positioning System) system, emitting radio signals received by the tag. The system can function in (TWR) Two-Way Ranging, TDoA 2 (Time Difference of Arrival 2) or TDoA 3 (Time Difference of Arrival 3) modes. The anchors and tags exchange radio signals. The system measures the time it takes for signals to travel between anchors and tags, which allows it to calculate distances. Using the distances from multiple anchors, the tag on the Crazyflie calculates its 3D position. This happens onboard without the need for an external computer. With precise position information, the Crazyflie can perform autonomous flight tasks, such as following waypoints, navigating mazes, or even flying in formations with other Crazyflies.

The lighthouse system relies on Lighthouse base stations. They emit sweeping infrared laser beams across the room. The Crazyflie needs the Lighthouse positioning deck attached. This deck has photodiodes that detect the laser sweeps from the base stations. When the laser beams hit the photodiodes on the Crazyflie, the system measures the precise angle between the base station and the receiver. By having multiple receivers on the Lighthouse deck (usually four), the Crazyflie can calculate its position and orientation with respect to the base stations. Like the Loco system, this calculation also happens onboard the Crazyflie, eliminating the need for communication with an external computer. However, both these systems do not come standard with the platform and must be purchased separately. There are also considerations such as line of sight issues, limited range, and need for calibration. Both active and passive light beacons face these issues. Figure 4 shows a common issue with beacons. Beacon 2 is behind an obstacle and cannot transmit information to the UAV (shown by the green sphere). Beacons 1 and 3 transmit information to the UAV; however, if the UAV moves farther out of the LoS (Line of Sight) of Beacon 3, it may lose access to the information that Beacon 3 is transmitting.

Figure 5 depicts an illustration of the lighthouse system that Crazyflie uses for positioning.

Despite lacking a GNSS receiver, there is a novel workaround for providing Tello drones with local positioning and global GNSS coordinates. This method is relatively cost-effective but has limitations in terms of range. The DJI Tello comes with standard ground mission pads that the downward-pointing sensor can recognize. The Tello uses these pads and the Visual Positioning System to localize and orient itself. By placing these pads on pre-decided local or staked GNSS global coordinates, it is possible to use the tellopy library and create an overlay map of the coordinates and the mission pad locations on the controlling computer. Since each pad is uniquely designated in the created map, once the Tello recognizes the mission pad, the information can be compared to the coordinate overlay map, and the drone gets the coordinate information specific to the recognized mission pad. Figure 6 shows the workflow for this described scenario.

2.3. Capability and Physical Facts

Table 4 presents a comparison of the physical characteristics and capabilities of the three platforms. An important ability in these EDU platforms is swarm support. UAV swarms consist of multiple UAV agents working in tandem, communicating with each other, often through an established hierarchy [15]. UAV swarms are increasingly used in applications to enhance the temporal efficiency and data collection capabilities of drones. Figure 7 shows two agents from a Robolink and Tello swarm working in tandem to locate a missing agent from the swarm in an indoor environment [11,12].

While the Tello and Robolink use downward-facing sensors to support physical ground-based markers, the Crazyflie does not have this as standard out-of-the-box support.

2.4. Usage and Support

Table 5 addresses the usage and support characteristics of the three platforms. The Tello and the Crazyflie have mobile applications to control the drones. The CoDrone platform does not. The presence of an official mobile app platform provides a convenience factor for drone usage. It also provides the user with additional ways to control the drone if the controller is unavailable. Cost-wise, the Tello seems to have chosen not to include the controller as a standard with the drone, instead pushing the user to rely on their mobile computing devices to control the drone. Several simulation platforms have also incorporated the ability to simulate popular drone platforms in simulation scenarios, making it easier to conduct experiments on the drones in simulation before they are tried on the hardware platform. This reduces the risk of failure and keeps costs in check. There is no direct simulator platform support for the Tello and the Robolink. Commercial versions of other DJI drones, such as the DJI Mavic, can be simulated on popular platforms such as the Webots simulator (link: https://cyberbotics.com). The Crazyflie is available as a drag-and-drop module in Webots. Figure 8 shows the Crazyflie and the DJI Mavic performing in the Webots simulation platform.

The UAV model in the popular CoppeliaSim platform closely mimics the DJI Tello design. Since CoppeliaSim allows modification of material properties such as material mass and sensor parameters, it can be modified to simulate the DJI Tello. Figure 9 shows a general quadcopter whose in-simulation model characteristics, such as mass, thrust, and sensor suite, were modified to be similar to the Tello.

Drone Blocks also offers a convenient simulation platform along with their block code support. Students and educators can now control the Crazyflie directly and code for it in the platform. The side-by-side simulator also offers a risk-free way to observe code directly in action on a simulator before being deployed on the actual drone. Figure 10 shows a screen capture of the Drone Blocks platform and the simulator. Figure 11 shows the Robolink CoDrone connected to a PC running Blockly and executing a mission.

Regarding ROS support, the CoDrone platform is not officially compatible with ROS. Tello EDU can be easily integrated into the ROS ecosystem. Popular packages and strong community support are available for both ROS1 and ROS2. Some packages have even succeeded in extending the capabilities of the drone, such as allowing Visual SLAM capabilities for mapping indoor environments. Crazyflie is one of the most popular platforms for ROS development in research and education. The official package “crazyswarm” has several features, such as drivers for communication with the Crazyflie, simulation tools for testing algorithms in a virtual environment, demo applications showcasing various functionalities, and support for swarm behavior, allowing you to control multiple Crazyflies simultaneously.

2.5. Use Cases and Suggested Integrations in Research.

The potential of these low-cost drone platforms has reached beyond the classroom into the domain of research applications. A brief literature review conducted during this study highlighted their use in four significant use case scenarios.

Figure 12 shows the four possible use case scenarios of educational drone platforms in research applications.

Figure 13 shows two common uses of educational drone platforms in current research. Table 6 and Table 7 list current literature that uses the Crazyflie and the DJI Tello platforms and outline the use case scenarios for each. Additional applications can be demonstrated using these platforms. Some future suggested use cases include using them as Blockchain-backed data collection and storage platforms [16], using LLM as a control structure for autonomous flight [17], and as agents of large-scale hierarchical swarms [18].

Due to its focus being more on classroom education and less on research applications, very little research exists that uses the Robolink EDU platform. However, the newer revised version of the platform shows promising capabilities that have potential in application use cases similar to the other two platforms.

2.6. Key Takeaways

Schools and universities are increasingly incorporating drones into STEM curricula to make learning interactive and relevant to 21st-century careers, particularly in robotics, automation, and data analytics. Moreover, drones are an accessible platform for introducing emerging technologies like artificial intelligence and machine learning through applications like autonomous navigation and object recognition. The potential of UAVs in education is vast, offering a dynamic approach to skill-building and knowledge acquisition that prepares students for the evolving demands of modern industries. Conceptual knowledge of aerodynamics, GNSS, inertial navigation, and electronics is communicated effectively through experiments and demonstrations on UAV platforms. Additionally, creative applications such as aerial photography and videography encourage students to think critically about storytelling and visual design.

An additional topic to be explored is the proliferation of complete DIY packages that involve using open-source files to make/print 3D aircraft models and individually source the electronic components to create modular drones from scratch. However, this approach is more hands-on and involves considerably more effort.

Readily available, low-cost educational platforms, like those discussed in this study, form a crucial bridge between higher-end commercial drones and DIY package drones. They offer a practical outlet for students to explore the domain before they venture further into it as research or a creative pursuit.