How FPV Drones Track Moving Targets: Technologies, Capabilities, and Future Prospects

Moving objects are among the most challenging scenarios for unmanned aerial vehicles (UAVs). Unlike a stationary target, a vehicle, another drone, or any other moving object constantly changes its position, speed, and direction.

That is why modern FPV drones and other UAVs are increasingly using computer vision, automated tracking, and motion prediction technologies.

Research in UAV tracking shows that key challenges include rapid changes in an object’s scale within the frame, camera movement, image blur, object occlusion by other elements, and the need to operate in real time.

How a Drone Detects a Moving Object

In the simplest scenario, the operator views the video feed from the onboard camera and manually controls the flight.

In more advanced systems, some of these tasks can be delegated to an onboard computer. The system analyzes the video, identifies an object within the frame, and attempts to maintain its position as it moves.

This process can generally be divided into several stages:

capturing an image from the camera;
detecting the object;
determining its position within the frame;
tracking changes in its position;
predicting its subsequent movement;
adjusting the flight trajectory.

Similar approaches are used not only in the defense sector but also in traffic monitoring, search-and-rescue operations, robotics, and other autonomous navigation systems.

The Role of Computer Vision

Computer vision enables a UAV to analyze video without requiring constant manual control of every element of its trajectory.

Modern algorithms can use:

neural networks for object detection;
visual tracking algorithms;
motion estimation;
sequential frame analysis;
object re-identification after partial occlusion.

Maintaining the identity of an object is particularly important. If an object briefly disappears behind an obstacle or changes its scale within the frame, the system needs to be able to identify and track the same object again.

Modern research in this field considers Siamese networks, Transformer architectures, and other machine-learning methods for this purpose.

Why Tracking Moving Objects Is More Difficult

Changes in Speed

An object may accelerate, slow down, or change direction. In such cases, simply predicting its movement based on its previous trajectory may not be sufficient.

Camera Movement

An FPV drone is constantly changing its own position. Therefore, the algorithm needs to distinguish between the movement of the object and the movement of the platform itself.

Image Blur

During high-speed flight or sudden maneuvers, the video image may become less clear. This makes it more difficult for computer vision algorithms to accurately identify and track objects.

Obstacles

Trees, buildings, vehicles, and other objects can partially or completely obscure the object being tracked.

Lighting Conditions

Nighttime, high-contrast lighting, fog, and other environmental conditions can also affect the quality of visual analysis.

Automated Tracking and Autonomy

One of the promising directions in the development of FPV platforms is the automation of certain operator tasks.

Modern systems can combine camera data with information about the movement of the aircraft itself. For example, research published in 2026 demonstrates approaches in which telemetry data is combined with computer vision to improve tracking stability during high-speed movement of an FPV platform.

This is particularly important for systems with limited computing resources, where algorithms need to operate virtually in real time.

The Importance of Autonomous Technologies

Autonomy is gradually becoming one of the key directions in the development of unmanned systems.

Depending on the platform architecture, automation may involve:

flight stabilization;
navigation;
object detection;
object tracking;
movement prediction;
trajectory adjustment.

At the same time, fully autonomous tracking remains a challenging task. Algorithms must operate in unpredictable environments, under unstable lighting conditions, in the presence of obstacles, and during sudden trajectory changes.

FPV Drones and Fiber-Optic Control Systems

Another important direction in the development of modern FPV platforms is the use of fiber-optic communication links.

Such systems can reduce reliance on traditional radio-control links. Research into the use of FPV systems in modern conflicts also identifies fiber-optic communication as one of the technological directions in the evolution of these platforms.

However, a fiber-optic link does not solve every challenge associated with autonomy. In complex scenarios, video-feed quality, navigation reliability, onboard systems, and the ability of algorithms to analyze a dynamic environment remain critical factors.

The Future of FPV Technologies

The further development of FPV drones is not limited to increasing flight speed or range.

Key development areas include:

Computer Vision. Algorithms are becoming increasingly capable of working with smaller objects and more complex visual environments.
Edge AI. Processing information directly onboard reduces dependence on external computing systems.
Multimodal Sensors. Combining data from different sources can improve system stability and reliability.
Motion Prediction. Algorithms are moving beyond simply tracking an object’s current position toward estimating its potential future trajectory.
Autonomous Navigation. UAVs are gaining greater capabilities for making navigation decisions independently.

Conclusion

The ability of an FPV drone to operate with moving objects depends not on a single component, but on an entire combination of technologies, including the camera, stabilization system, onboard computer, computer vision algorithms, navigation, and software.

The transition from simple remote control to intelligent tracking and autonomous navigation is becoming one of the most important directions in the development of modern unmanned systems.