DEVELOPING A FORMALISM FOR GIBSON’S AFFORDANCES USING COLORED PETRI NETS
Open Access
- Author:
- Thiruvengada Ramanujam, Hari H S
- Graduate Program:
- Industrial Engineering
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- January 31, 2007
- Committee Members:
- Ling Rothrock, Committee Chair/Co-Chair
Gwendolyn E Campbell, Committee Member
Richard Allen Wysk, Committee Member
M Jeya Chandra, Committee Member
Sean N Brennan, Committee Member - Keywords:
- Human Behavior Modeling
Performance Evaluation
Driver Behavior
CPN
Colored Petri Net
Affordance
Simulation - Abstract:
- Gibson (1979/1986) proposed affordance theory to represent and model what the environment offers an animal for good or ill. Since its inception by Gibson, affordance theory has undergone several refinements. A few affordance theory-based formalisms are reviewed in this proposal to demonstrate their potential advantages and disadvantages and to motivate an overarching formalism to model problems within dynamic environments. The purpose of this research is to provide a computational formalism for Gibson’s affordance theory based on characteristics of dynamic environments to include concurrency, stochasticity and spatio-temporality. A Colored Petri Net (CPN)-based model is proposed as a suitable instrument for developing this formalism. A mathematical model, graphical representation and computational model for this CPN model is developed within the context of a driving problem. The affordances offered by this driving environment are analogous to those offered by a set of highway lanes. A formative analysis technique is also introduced along with an overall data analysis procedure to analyze the precision of the actualized actions and the niche of lane affordances that become available to the driver within the highway lane-driver system. An empirical study was conducted using a team of two expert drivers to elicit various behaviors that would help resolve the precision of the CPN model. Four output metrics were defined that represent the deviation between the empirical human performance and model predicted data: lane position, turn direction of the subject driver’s vehicle, time taken by the subject driver to move from the starting lane to the exit lane and the total utilization of the exit lane by the subject driver. The significance of the results is then explained with reference to research implications and future work.