The recognition and application of light language in unmanned vehicle environment understanding

Environment understanding is the prerequisite for safe driving of unmanned vehicle. However, most unmanned driving environment understanding researches neglect to use communication information between vehicles to predict the behaviors of vehicles surrounding unmanned vehicle. The paper proposes to use light language, the main way that vehicles exchange information, to predict behaviors of surrounding vehicles. Based on vehicle detection, the paper studies how to detect the position and status of light of vehicles. Then, combined with environment information, light language model database is established to recognition light language. Finally, we apply light language to achieve control strategy for unmanned vehicle. The application of light language will afford more time to unmanned vehicle to response the behaviors of surrounding vehicles, and ensure unmanned vehicles against traffic accidents.