Open Access Dissertation
Date of Award
Industrial and Manufacturing Engineering
Richard Darin Ellis
Robotic microsurgery provides many advantages for surgical operations, including tremor filtration, an increase in dexterity, and smaller incisions. There is a growing need for a task analyses on robotic laparoscopic operations to understand better the tasks involved in robotic microsurgery cases. A few research groups have conducted task observations to help systems automatically identify surgeon skill based on task execution. Their gesture analyses, however, lacked depth and their class libraries were composed of ambiguous groupings of gestures that did not share contextual similarities.
A Hierarchical Task Analysis was performed on a four-throw suturing task using a robotic microsurgical platform. Three skill levels were studied: attending surgeons, residents, and naïve participants. From this task analysis, a subtask library was created. The Hierarchical Task Analysis subtask library, a computer system was created that accurately identified surgeon subtasks based on surgeon hand gestures. An automatic classifier was trained on the subtasks identified during the Hierarchical Task Analysis of a four-throw suturing task and the motion signature recorded during task performance. Using principal component analysis and a J48 decision tree classifier, an average individual classification accuracy of 94.56% was achieved.
This research lays the foundation for accurate and meaningful autonomous computer assistance in a surgical arena by creating a gesture library from a detailed Hierarchical Task Analysis. The results of this research will improve the surgeon-robot interface and enhance surgery performance. The classes used will eliminate human machine miscommunication by using an understandable and structured class library based on a Hierarchical Task Analysis. By enabling a robot to understand surgeon actions, intelligent contextual-based assistance could be provide to the surgeon by the robot.
Limitations of this research included the small participant sample size used for this research which resulted in high subtask execution variability. Future work will include a larger participant population to address this limitation. Additionally, a Hidden Markov Model will be incorporated into the classification process to help increase the classification accuracy. Finally, a closer investigation of vestigial techniques will be conducted to study the effect of past learned laparoscopic techniques, which are no longer necessary in the robotic-assisted laparoscopic surgery arena.
Golenberg, Lavie Pinchas, "Task Analysis, Modeling, And Automatic Identification Of Elemental Tasks In Robot-Assisted Laparoscopic Surgery" (2010). Wayne State University Dissertations. 140.