Accurate road information warning framework for intelligent vehicles based on resource allocation

Abstract

The main goal of the road information warning is to notify intelligent vehicles of road situations in advance so that they have sufficient time to correctly react and avoid road hazards and accidents. However, an intelligent vehicle has limited visual angles and resources, so it cannot produce accurate road information alone. Road information is classified into emergency information such as accidents and non-emergency information such as traffic conditions. Emergency information is issued via flooding where information fusion is not allowed. This flooding results in huge information redundancy and inaccurate road information. Non-emergency information is retrieved by the request–response model where only the requestor can access road information. This model leads to inefficiency of information retrieval because intelligent vehicles need to share road information. Furthermore, road information retrieval may fail if vehicles without abundant resources participate in delivering road information. Taking into account these issues, we propose a road information warning framework for intelligent vehicles based on resource allocation. To generate accurate road information, vehicles exploit local clouds to share resources and cooperate to produce road information. To enhance emergency information accuracy and suppress information redundancy, the naming and resource allocation mechanisms are proposed to achieve information fusion and improve efficiency of issuing emergency information, respectively. To achieve non-emergency information sharing, the information aggregation is proposed so that intelligent vehicles can rapidly access information from the nearest provider. The framework is evaluated and the experimental results verify its advantages.