Abstract
The key to an ultrasound-guided system is to accurately determine the spatial position of the surgical instrument. Traditional optical positioning devices often restrict a surgeon's movements and occupy significant operating room space. This paper presents a compact device designed to provide enhanced navigation information. Equipped with an angle sensor and a manual depth-adjustment button, the device measures the relative angle between a puncture needle and an ultrasound probe in real time. A custom two-stage filtering algorithm was developed to reduce signal noise. Results show significant improvements in signal smoothness and a low average point-to-point error of 3.53 pixels, demonstrating the device's potential to support precise surgical operations without the need for bulky external hardware.
Introduction
Ultrasound-guided surgery is a cornerstone of minimally invasive procedures, yet maintaining spatial awareness of the needle's path relative to the probe remains a significant challenge. Current "gold standard" solutions typically rely on optical tracking systems. These systems require mounting bulky targets on surgical tools and positioning binocular cameras around the operating room, which can obstruct the surgeon’s line-of-sight and consume valuable space.
To address these limitations, researchers have developed a mechatronic attachment that "shrinks" the navigation system. By integrating sensors directly onto the ultrasound probe, the device provides real-time guidance and depth estimation without the need for external tracking hardware.
System Design and Hardware Implementation
The proposed device is a 3D-printed mechatronic sleeve designed to fit standard ultrasound probes. Its primary function is to measure the relative tilt between the puncture needle and the probe using a high-precision RDC501051A resistive angle sensor.
The system's core is an STM32F103C8T6 microcontroller, which processes data from the angle sensor via a 16-bit ADS1110 analog-to-digital converter (ADC). This architecture ensures that even minute changes in the needle's orientation are captured with high fidelity, allowing for precise tracking of the instrument's trajectory.
Signal Processing and Depth Estimation
A critical challenge in medical mechatronics is signal noise caused by electromagnetic interference and mechanical vibrations. To ensure a stable visual guide for the surgeon, a two-stage hybrid software filter was implemented:
- Windowed Mean Filter: This stage removes extreme values from a 10-point window and calculates the mean to eliminate transient spikes.
- Standard Deviation Outlier Removal: A larger 100-point window is used to identify and strip outliers based on standard deviation, followed by a final mean filter for maximum smoothness.
Furthermore, the device includes a physical button that allows surgeons to adjust a "virtual guide line" on the ultrasound display. This enables pre-puncture depth estimation, established through a polynomial fitting model ($y = 16.32x - 1.074$), allowing for a visual "dry run" of the needle path before the actual procedure begins.
Performance Evaluation and Results
The device was evaluated using a breast body model and a dedicated experimental bench. The filtering algorithm proved highly effective, reducing the Signal Smoothness metric from 0.036989 to 0.0010376 and cutting the Count of Local Extrema (sharp, erratic signal jumps) from 251 to 135.
In accuracy tests, the average point-to-point error between the predicted guide line and the actual needle tip was found to be 3.53 pixels. Depth measurement precision was equally impressive, with an average error of just 0.216 mm and a near-perfect R-squared value of 1 for the polynomial model, indicating a highly reliable correlation between the device's sensor data and physical depth.
Conclusion
The developed guidance device and filtering algorithm provide a compact, accurate, and non-obtrusive alternative to traditional optical tracking systems. By moving the navigation intelligence directly onto the probe, the system supports surgical precision while streamlining the operating room environment. Future work will focus on automating depth measurement using magnetic sensors and increasing the sampling rate to further enhance real-time responsiveness.
Summary updated