My senior design project focused on creating a device that measures advanced exercise metrics such as barbell velocity, acceleration, and tilt. These are important for optimizing strength training and rehabilitation but are not captured by traditional methods. To solve this, I developed a smart barbell attachment: a lightweight, wireless sensor system that uses an accelerometer and gyroscope connected to a microcontroller. This system records real-time motion data, which is then processed using algorithms to provide detailed feedback on lifting performance. Through testing, the device successfully distinguished between light and heavy lifts, detected points where lifters tend to struggle, and identified barbell tilts that may indicate muscle imbalances or technique flaws. These results validated that the system could provide athletes, coaches, and physical therapists with useful, actionable data that normally goes unseen during workouts.
I became interested in this research area because of my background in powerlifting and my desire to explore ways to measure lifting performance more precisely. I was motivated by the idea that deeper, more accurate measurements could not only improve athletic outcomes but also prevent injury by identifying poor form early. During my initial research, I reviewed several commercial products and found that while some measured bar velocity, none offered comprehensive tilt angle tracking or gave users access to raw sensor data for deeper analysis. That realization made me confident there was a meaningful opportunity for innovation.
Going into the project, I expected the capstone to be centered on building hardware and running some tests to demonstrate it worked. In reality, the experience was much richer and more complex. It required me to study and apply advanced techniques like sensor fusion through Kalman filtering, coordinate transformations using rotation matrices, and digital signal processing in MATLAB to clean up noisy data. The project became an interdisciplinary challenge, bringing together hardware design, embedded programming, algorithm development, and biomechanical analysis.
One of the hardest challenges I encountered was dealing with sensor imperfections. Accelerometers tend to suffer from noise, while gyroscopes are prone to drift, making raw measurements unreliable. Combining these two sources through sensor fusion to get stable, accurate readings turned out to be a detailed and technical process. Another major challenge was designing an algorithm to detect exercise repetitions based on velocity data. Lifting movements in real life are inconsistent. Pauses, varying tempos, and small deviations made simple threshold detection unreliable. I had to build a finite state machine (FSM) that could adapt to these irregularities without miscounting repetitions. On the other hand, some aspects went more smoothly. Designing and assembling the physical prototype, including the 3D-printed case, was relatively straightforward and satisfying. I was also surprised by how much trial and error went into tuning filters. Extracting clean velocity signals required carefully balancing filter settings to reduce noise without removing valid movement data.
If I were to continue this work, I would first add Bluetooth capability to the device.
Currently, the prototype logs data onto a micro-SD card, which means users have to manually transfer the data for analysis. Wireless streaming to a mobile app would make the system much more convenient and practical. This would also open the door to real-time feedback, such as audio or visual cues when a user’s form begins to break down during a set. Another area for expansion would be adapting the data analysis algorithms to run directly on the device or in the app, so that lifters receive instant feedback rather than having to review reports afterward.
Finally, I would test the device with a broader group of users with different body types and lifting styles to make the repetition detection algorithm more robust and adaptable to diverse patterns. The most important takeaway from this project is a deeper appreciation for the challenge of capturing and analyzing human movement in a reliable way. I developed confidence in building systems from start to finish, from the idea stage through hardware assembly and software programming to producing useful feedback for users. This process also taught me how to comprehensively document my designs and create professional-quality figures and schematics that clearly communicate technical ideas.
No comments:
Post a Comment