A method of vehicle-infrastructure cooperative perception based vehicle state information fusion using improved kalman filter

For the purpose of overcoming the technical bottlenecks and limitations of autonomous vehicles on the information perception, and improving the sensing range and performance of vehicle driving environment and traffic information, a framework of vehicle-infrastructure cooperative perception for the Cooperative Automated Driving System is proposed in this paper. Taking the vehicle state information as an example, it also introduced a calculation method of data fusion for vehicle-infrastructure cooperative perception. Besides, considering that the intelligent roadside equipment may appear short-term sensing failure, the proposed method improved the traditional Kalman Filter to output position information even when the roadside fails. Compared with the vehicle-only perception, the simulation experiments verified that the proposed method could improve the average positioning accuracy under the normal condition and the intelligent roadside failure by 18% and 19%, respectively. The proposed framework provided a solution for coordinating and fusing perception intelligence and functions between connected automated vehicles, intelligent infrastructure and intelligent control system. The proposed improved Kalman Filter method provides flexible strategies for practical application.

Automated summary

The paper proposes a framework of vehicle-infrastructure cooperative perception for the Cooperative Automated Driving System, along with a method of data fusion for vehicle-infrastructure cooperative perception. The proposed method can output position information even when the intelligent roadside equipment fails, and it improves the positioning accuracy under normal conditions and roadside failure.