Validation of a Signal Processing and Sensor Data Fusion Pipeline for Cochlear Implant Insertion Force Measurement

Introduction: Insertion forces acting during cochlear implant surgery have been associated with intracochlear trauma and loss of residual hearing, but are almost imperceptible. To enable intraoperative insertion force measurement, our group developed the Forception Tool, a handheld device integrating a force sensor and an inertial measurement unit (IMU) for correcting gravity-related bias in force signals. While the tool and its data acquisition pipeline are established, the signal processing and sensor fusion validation required for reliable force interpretation are yet to be performed. Methods: Sensor-specific noise was characterized, and low-pass filters were designed for the force and IMU accelerometer signals. The existing pipeline was then extended to acquire, filter, and fuse the sensor data, yielding a filtered gravity-compensated force signal. Filter performance was evaluated using data from 30 automated linear insertions with the tool, employing two metrics: Peak Preservation Ratio (PPR) and Residual Root Mean Square (Residual RMS). To verify synchronicity in the sensor fusion, 20 drop experiments were conducted, and the onset times of the sensor responses were compared. Results: With filtering, relevant force peaks were preserved, achieving a mean PPR of 99.81%±0.02% and a Residual RMS of 0.26mN±0.03mN. Temporal synchronization between the streams was verified with a mean offset of 11 ms ± 4 ms. Conclusion: The implemented pipeline successfully integrates and validates signal processing steps and confirms temporal alignment between the sensor data streams. The results demonstrate the reliability of measurements with the Forception Tool and pave the way for real-time force feedback to surgeons.