SBIR-STTR Award

We propose to develop a knowledge base of driving hazards that may affect drivers aged 20-35 years. The knowledge base will include information about the hazards, associated risk of crash involvement, and practical behaviors for mitigating this risk. The knowledge based will be developed through a combination of exploratory data collection and analysis of existing data sources (analysis of the 100 car data set, critical incident interviews) and more experimental data collection using a driving simulator to establish the impact of internal and external hazards. This combination of methods will increase the accuracy of knowledge about hazards, and provide converging evidence of the importance of the identified hazards, underlying causes, and potential responses to mitigate risk. We conceive of driving skill, and especially risk assessment, as a form of tacit knowledge. As such, a training method effective in increasing tacit knowledge, namely scenario-based training, will be used to deliver risk assessment training in a web-based environment. Students will respond to the scenarios, and the system will provide immediate assessment and feedback on student responses.

Keywords:
Driving, Driving Wisdom, Driving Safety, Tacit Knowledge, Latent Semantic Analysis, Driving Risk Assessment, Critical Incident Interviews

Over 50% of accidental deaths in the Army during 2006 occurred in privately owned vehicles (POV) the most attributable to one source. As a result, the Army has been actively pursuing ways of reducing POV deaths and injuries among Soldiers. In the Phase I effort we designed an effective web-based training program to improve driver safety by training drivers to understand hazards, more accurately assess their risk associated with hazards, and take the appropriate actions in response to the presence of hazards. We have four primary goals for the Phase II effort. The first is to continue to develop the database of driving hazards, associated risk, and practical strategies, both as a way of developing the additional scenarios for the full prototype, and as a way of including less well-understood hazards such as tactical driving. Second, we intend to bring the demonstration prototype developed in Phase I to a complete prototype that is usable by the Army. The third goal is to provide some evaluation of the effectiveness of the training system. Finally, our fourth goal is to explore ways of extending the prototype and increasing the breadth of training.

Keywords:
Driving Safety, Tacit Knowledge, Scenario-Based Training, Machine Learning, Driving Simulator