Running is one of the most accessible forms of exercise, requiring nothing more than a pair of shoes and a path. Yet for many, the simple act of jogging can be interrupted by a sudden struggle to breathe, a sensation that halts momentum and dampens the joy of the activity. This respiratory distress often stems from a disconnect between how a person moves and how they breathe. In the world of sports science, there is a known rhythm where the body naturally seeks to lock its breathing pattern to its footsteps, a synchronization that can make running feel easier and more efficient. However, finding and maintaining this rhythm is difficult for beginners, and traditional methods of guiding runners often rely on looking at what happened in the past. If a runner suddenly changes their pace or their breathing pattern, systems that react to old data are too slow to help, offering advice that arrives a moment too late to be useful.

To solve this problem, researchers at Salzburg Research and the University of Salzburg have developed a new approach called BPqu. This system is designed to be a real-time guide that predicts a runner's breathing before it fully happens, rather than just reacting to it after the fact. By placing a small, powerful computer directly onto a wearable sensor attached to a runner's chest, the system can analyze breathing and movement data instantly. The goal is to provide immediate, intuitive feedback that helps runners find their natural rhythm without the lag that frustrates users of older technologies. This work represents a shift from observing the past to anticipating the present, aiming to remove the cognitive and emotional barriers that keep people from enjoying endurance running.

The core challenge the team faced was the delay inherent in most digital systems. When a computer processes information, it takes time to collect data, crunch the numbers, and send a signal. In a running scenario, even a fraction of a second of delay can mean the difference between a helpful cue and a confusing one. If a runner is told to exhale after they have already started to inhale, the guidance is useless. The researchers needed a way to determine exactly where a runner is in their breathing cycle at this very moment, with zero delay. They also needed to predict the next step in the cycle to ensure the guidance arrives exactly when it is needed. To achieve this, they turned to a type of artificial intelligence known as deep learning, which can recognize complex patterns in data, but they had to shrink this technology down to fit inside a tiny, battery-powered device worn on the body.

The researchers built a custom sensor that measures both the pressure changes in the chest during breathing and the movement of the body during each stride. This sensor is connected to a small microcontroller, a chip powerful enough to run complex calculations but small enough to be embedded in clothing. The team trained a computer model using data collected from nineteen young female runners. These participants ran while wearing the sensor, and their breathing was carefully monitored to create a library of patterns. The model learned to recognize the subtle signals that indicate when a runner is inhaling, when they are exhaling, and exactly when the switch between the two occurs. Crucially, the model was designed to look at the relationship between the breath and the steps, understanding that the two are often linked in a specific rhythm.

To make the system fast enough for real-time use, the researchers had to be clever about how they asked the computer to think. Instead of trying to guess the entire breathing curve at once, which can be messy and prone to error, they broke the problem into two simpler parts. The system estimates the general shape of the breathing cycle and separately identifies the exact moments when the breath switches from inhaling to exhaling. It then combines these two pieces of information to create a complete, accurate picture of the runner's current state. This method allowed the system to work with a very small amount of memory and processing power, making it possible to run entirely on the wearable device without needing to send data to a cloud server or a smartphone.

The results of this approach were impressive. When tested, the system could identify the exact moment a runner switched from inhaling to exhaling with an accuracy of 93.1%. It could also determine whether the runner was currently inhaling or exhaling with an accuracy of 85.6%. Perhaps most importantly, it could quantify the entire breathing phase with an accuracy of 81.5%, all without any measurable delay. This means the system is fast enough to keep up with a runner even if they suddenly speed up or change their breathing pattern. The researchers found that combining data from both the breathing sensor and the movement sensor worked better than using breathing data alone, as the movement data helped the system filter out noise caused by the jostling of running.

The study also explored how well the system performed under different conditions. It worked best when the runner was breathing in a steady, predictable rhythm, but it remained effective even when the breathing became irregular or when the runner was struggling to maintain a rhythm. The system was tested in three different modes: looking back at what just happened, looking at the current moment, and looking slightly ahead to predict the next moment. In all cases, it maintained high accuracy, proving that it could handle the dynamic nature of running. The researchers noted that while the system is highly accurate, it is not perfect, and there is still a small margin of error, particularly when breathing patterns are highly irregular. However, the speed of the system means that even with a small error, the feedback is delivered in time to be useful.

This work opens the door for a new kind of running experience. Imagine a runner who can receive a gentle, rhythmic cue that tells them exactly when to breathe, helping them find a flow that feels natural and effortless. Because the system works instantly, it can adapt to sudden changes in pace or terrain, offering guidance that feels intuitive rather than robotic. The researchers envision this technology being integrated into smart clothing, providing a seamless way to support runners who might otherwise give up due to breathlessness. By removing the frustration of poor breathing control, the technology aims to make running more enjoyable and accessible, helping people push past their limits and find a deeper connection with their physical activity.

The path forward involves testing this system in real-world running scenarios, outside the controlled environment of a laboratory. The researchers plan to see how runners react to the guidance in actual races or training sessions, and whether the system can truly help people overcome the cognitive and emotional barriers that make running difficult. They also hope to refine the system to work with a wider variety of users, including those with different fitness levels and breathing patterns. The ultimate goal is to create a tool that not only tracks performance but actively enhances the human experience of running, turning a struggle for breath into a moment of flow and enjoyment. This research demonstrates that by combining advanced computing with a deep understanding of human movement, it is possible to create technology that feels less like a machine and more like a helpful companion on the run.