A modular extreme learning machine with linguistic interpreter and accelerated chaotic distributor for evaluating the safety of robot maneuvers in laparoscopic surgery

作者：Mozaffari Ahmad; Behzadipour Saeed^*

来源：Neurocomputing, 2015, 151: 913-932.

DOI：10.1016/j.neucom.2014.10.003

• Laparoscopic surgery
• Medical robotics
• Soft tissue modeling
• Clustering
• System identification
• Fuzzy inference system

In this investigation, a systematic sequential intelligent system is proposed to provide the surgeon with an estimation of the state of the tool-tissue interaction force in laparoscopic surgery. To train the proposed intelligent system, a 3D model of an in vivo porcine liver was built for different probing tasks. To capture the required knowledge, three different geometric features, i.e. Y displacement of the nodes on the upper surface and slopes on the closest node to the deforming area of the upper edge in both X-Y and Z-Y planes, were extracted experimentally. The numerical simulations are conducted in three independent successive stages. At the first step, a well-known partition-based clustering technique called accelerated chaotic particle swarm optimization (ACPSO) is used to cluster the information of database into a number of partitions. Thereafter, a modular extreme learning machine (M-ELM) is used to model the characteristics of each cluster. Finally, the output of M-ELM is fed to a Mamdani fuzzy inference system (MFIS) to interpret the safety of robot maneuvers in laparoscopic surgery. The proposed intelligent framework is used for real-time applications so that the surgeon can adjust the movements of the robot to avoid operational hazards. Based on a rigor comparative study, it is indicated that not only the proposed intelligent technique can effectively handle the considered problem but also is a reliable alternative to physical sensors and measurement tools.

• 出版日期2015-3-5