Detecting Concentric Muscle Failure in Bench Press: A Bag of Patterns Approach using Wearable Data

As urban life becomes increasingly fast-paced, encouraging healthier habits and ensuring safety during physical activity are more important than ever. Sports safety is key to supporting athletes, fitness professionals, and everyday users by helping prevent injuries and optimizing training. One critical aspect is detecting concentric muscular failure (CMF), a point during resistance training, such as the bench press, when a person cannot lift the weight with the proper technique. Identifying CMF in real time allows for timely intervention and safer, more personalized workouts. In this work, we propose using the Bag of Patterns (BoP) technique to detect concentric muscular failure during bench press exercises from wearable data. Experiments conducted with data collected from 14 participants across two sessions of bench press exercise in real-world scenarios demonstrate that combining the BoP approach with logistic regression achieves an accuracy of 88 %, while using raw data with KNNDTW results in an accuracy of 84 %. These findings highlight the potential of wearable-based systems in improving exercise safety and performance.