Modeling the Effects of Time Lag in Virtual Reality (VR)-Based Haptic Surgical Simulator

Show full item record

Title: Modeling the Effects of Time Lag in Virtual Reality (VR)-Based Haptic Surgical Simulator
Author: Li, Yingjie
Advisors: Dr. Simon Hsiang, Committee Member
Yuan-Shin Lee, Committee Co-Chair
Dr. David B. Kaber, Committee Chair
Abstract: The goal of this research was to evaluate a virtual reality (VR)-based simulation of a surgical task by using a model of human motor-control behavior to describe performance. The study was to serve as a basis for formulating general simulator design guidelines. In VR-based simulator design, there are often tradeoffs made between simulation fidelity, or realism, and computational (rendering) load on system processors to ensure acceptable performance. The greater the rendering load, the slower the response of the simulation to user actions and the lower user perceptions of realism. One way to assess the level of fidelity of a VR simulation for surgical training is to determine whether user motor behavior in virtual tasks conforms to established models of real discrete movement task performance, such as Fitts' Law. The implications of computational load on simulator realism can be evaluated by introducing artificial time delays in system responsiveness to control interface actions and determining whether lag is a significant predictor of motor control performance degradations, as compared to no lag conditions. Such assessment can be made by using an extension to Fitts' Law. A secondary objective of this study was to determine whether the impact of computational load on the conformance of motor control behavior with Fitts' Law varied across surgical task complexity settings. In addition to using a Fitts' task to assess the VR simulation, a tracking task with artificial time delays was simulated to assess the level of fidelity of the VR simulation on continuous motor performance. A simple surgical simulator was developed for the study. User performance was tested in virtual tissue cutting (with a virtual scalpel) under different system lag conditions and settings of task difficulty. Four levels of lag were introduced in the response of the simulation to a haptic control device including 0 msec, 50 msec, 100 msec and 150 msec. Two levels of task complexity were investigated,as described by Fitts' index of difficulty (ID; 4 and 5). During training trials before the formal experiments, subjects were asked to repeat the cutting task with lag=0 and ID=4 until they were able to fully control the virtual scalpel. Under each combination of test conditions, the user was asked to use the stylus of a PHANTOM haptic device, control the virtual scalpel to reach a marked area on the tissue, cut to a desired depth, and cut along a marked path on the surface to an end point. The time and position of the virtual scalpel tip was recorded throughout the process. The root mean square error (RMSE) of scalpel position for the path tracking in the plane of the tissue surface and in the plane perpendicular to the surface was recorded in each test trial. Results on the Fitts' task (moving the scalpel to the marked tissue) revealed user performance to degrade with increasing task difficulty and time lag. In general, motor control behavior under the no lag condition was explained by Fitts' Law, supporting the fidelity of the simulation. An extended model of human motor performance in the surgical training simulation, incorporating a control lag parameter, appeared to be highly significant in explaining the data for the time delay conditions. An analysis of the interaction effect of task complexity and time delay revealed the impact of lag to vary across levels of ID, as observed in real discrete movement task performance. Tracking task results revealed an influence of parameter settings of the haptic control model on the VR application accuracy that interacted with the lag condition. This research demonstrated that motor behavior in a simple VR-based surgical simulation conforms to Fitts' Law across levels of ID. Consequently, if simulated surgical tasks impose high requirements on user response time, designers need to carefully control the computational load of a simulation and time lag. An extended Fitts' model was validated in this study, allowing one to predict the acceptable time lag in control if desired ranges of user response times and levels of surgical difficulty (described in terms of ID) can be specified. Lag appears to also have important subjective affect on the stability and perceived realism of a haptic model and VR simulators. The tracking results motivate additional research of haptic model parameters for VR to realistically simulate specific forms of tissue. Additional work is needed to determine design parameters to balance user feelings of realism with the need for system stability in the presence of time lags. In general, the findings of this work support predictions on human performance under various surgical simulator design conditions. They also facilitate a priori determination of whether certain VR simulation designs may be unacceptable from a training perspective.
Date: 2007-04-06
Degree: MS
Discipline: Industrial Engineering

Files in this item

Files Size Format View
etd.pdf 463.7Kb PDF View/Open

This item appears in the following Collection(s)

Show full item record