Towards a new era of surgery using intelligent robot

Authors

Ahmed Mahfouz, Department of Mechanical Engineering, Faculty of Engineering, Alexandria University, Alexandria, EgyptFollow
Hassan A. El Gamal, Department of Mechanical Engineering, Faculty of Engineering, Alexandria University, Alexandria, Egypt
T. Awad, Department of Mechanical Engineering, Faculty of Engineering, Alexandria University, Alexandria, Egypt
M.B. Badawi, Department of Mechanical Engineering, Faculty of Engineering, Alexandria University, Alexandria, Egypt

DOI

10.70259/engJER.2024.851856

Abstract

This research aims to overcome the difficulties of surgical tumor removal, especially in brain tumors that needs precision and time saving of the operations. The traditional surgical approach often is slow and targeting accurate manual cutting of the tumor, which is time-consuming and increase the risk to the patient. The error in cutting the deep layers of the tumor can occur, which increase the cost and dangerous for the patient.

To overcome these challenges, a proposed solution involves designing a robot that can quickly and accurately remove tumors by following their contours as it comes from the medical imaging data, such as DICOM-formatted images from CT, MRI, and ultrasound scans. This intelligent robot is trained by employing artificial neural networks. A suggested robot has 7 degrees of freedom, and its workspace is determined various positions of its links using DH tables and feed-forward kinematics. To achieve the desired movements, more than on artificial neural network is applied such as generalized Regression(GRNN), Radial Basis (RBNN), and Feed Forward (FFNN). These are trained using a dataset consisting of more than 200000 workspace points of the robot's end effector for inverse kinematics.

After training, the Feed-Forward NN is identified as the most accurate, this network is applied to control the robot's links motion during surgery. As a prework for the surgical procedure, the patient's medical images are digitized. The obtained data fed as the input for the trained FFNN, which determines the positions of the robot's links needed to accurately follow the tumor contour. By doing so, the robot can swiftly navigate and remove the tumor while maintaining precision and safety. This reduces the time of the surgery and minimizes the risks associated.

To achieve high accuracy more workspace points must be included in the training process of the NN. This increase the computational load, memory usage and extended time in training. this can be only achieved by using super computers.

Recommended Citation

Mahfouz, Ahmed; El Gamal, Hassan A.; Awad, T.; and Badawi, M.B. (2024) "Towards a new era of surgery using intelligent robot," Journal of Engineering Research: Vol. 8: Iss. 5, Article 9.
DOI: 10.70259/engJER.2024.851856
Available at: https://digitalcommons.aaru.edu.jo/erjeng/vol8/iss5/9