SAVED

Title: SAVED: System for Adapting Vehicle dynamic parameters to the driving Environment and Driver capabilities

Starting Date: 1st September 2009
Expected Completion Date: 31st August 2012

Principal Investigator: Professor José Manuel Viegas

Executive Coordinator: Sílvia Shrubsall

Institutions/Research Centers involved: Instituto Superior Técnico, Faculdade de Ciências e Tecnologia da Universidade de Coimbra (FCT/UC), Universidade do Minho 

Keywords: Intelligent Transport Systems (ITS), Intelligent Vehicle Safety Systems, Road Safety

Abstract:

Over the past 5 to 8 years, the number of accidents and fatalities has become constant, or even slightly increasing, following a 10 to 15 year period of a steady decrease of those indicators in countries with long road safety tradition. This suggests that the traditional road safety measures, including existing intelligent vehicle measures might have reached their limit of capability of significantly decreasing the number and severity of accidents. In Portugal, despite the rapid mobilization, the number of road accidents has lately decreased considerably. It is, however, expected that the same pattern as in the more developed countries is going to be repeated on a shorter timescale. Human factors contribute to over 95% of the road accidents and the driving task does not continually present the same degree of difficulty. SAVED, in the area of Intelligent Transportation Systems, seeks to employ innovative methods to provide a lasting significant improvement to road safety.

Central to the research proposal is the fact that drivers' competence is not constant over time and that the road environment, as well as external conditions, like the weather, pose different degrees of difficulty to the driving task. A structured strategy was conceived which will continuously assess those competencies and surrounding conditions and then both correct the car attributes and limit the degree of freedom of the driver to compensate for real-time shortcomings. We aim at expanding this concept followed by its development into a three-unit prototype, which will then be tested in a laboratory with a simulator, and subsequently, yet beyond the scope of the current proposal, to have a pre-industry application. The proposed system is primarily aimed at road safety, but is expected to have social and political impacts by, for example, influencing mobility choice, affecting social integration and contributing to law enforcement. The overall outcome of the system, including its acceptability, cost and real reduction of accident risks (and insurance premia), will determine its voluntary or mandatory basis of implementation by regulatory agencies and by the industry.

Overall Objectives:

The main aim of this study is to contribute to the improvement of road safety through the application of Intelligent Vehicle Combined Passive and Active Safety Systems. Its main objective is to recommend a system to adjust the vehicle's dynamic attributes to the driver's state and driving circumstances, which will ultimately grow into a technological device. As such, this work proposes to contribute to developing the specifications of a three-unit (Sensor Module, Risk Profiling System, and Control Tool) in-vehicle technological system which will continuously assess drivers' competencies and surrounding conditions, and automatically adjust the action space of the driver to preserve the desired safety levels. Figure 1 outlines the proposed system and underlying approach. Four initial objectives have therefore been identified:

• To list drivers' (permanent and temporary) limitations and driving circumstances which have an impact on road accident involvement;
• To develop a procedure to identify which aspects of risk perception and vehicle control could be affected by each of the limiting factors;
• To specify information or vehicle dynamics adjustments for those limitations. If amendment for one or more limitations is not possible or does not substantially reduce the probability and severity of accidents, to identify a list of restrictions of the driving task in one or more dimensions (e.g. speed, longitudinal and lateral acceleration, and distance to the vehicle ahead); and
• To develop a business model.

Highlights to date:

An initial computer model for about 500 meters of a non-urban highway and 80 cars driven have been developed using Agents in a MultiAgent-Based program named AnyLogic. This will be scaled up by using the outputs of this sort of models in traffic programs, like AIMSUN.