Training Simulations for Crime Risk Assessment

Abstract

The purpose of this paper is to review training simulations for crime risk assessment and to discuss the system architecture of a training simulation (RiskMan). Computer aided training systems offer a flexible and cost-effective method for learning new skills. The satisfactory management of risk situations involves risk identification, the development of risk handling strategies and plans and the conduct and monitoring of those plans. Recognising the importance of tacit knowledge, scenario-based training has gained importance in recent years. For example, USA General Accounting Office (GAO) released a report on Homeland security titled Risk Management approach can guide preparedness efforts. This report provides a list of risk assessment measurements for contingency plan development and a matrix for risk-based scenario development. In this paper, integrating traditional scenario-based training with desktop VR systems using game engineering, we investigate how a virtual reality training system, which draws on research in the areas of computer games, knowledge acquisition, agent technology and natural language processing, can provide a safe learning experience to assist acquisition of the necessary tacit knowledge. The aim of RiskMan is to train police officers to handle high-risk situations. RiskMan is an ARC Discovery project carried out by the Department of Computing in Macquarie University. RiskMan has been developed using a very-high level scripting language of a game engine, Unreal Tournament 2004. It is composed of modules such as a Scenario-based Expert System, a Narrative Engine, a Game Engine, and a CAD package. RiskMan uses socket connections to feed information between the Narrative Engine and Sim Master to Unreal Tournament Game Engine (UT2004)