An Integrated Dual Event-Driven Approach for Motion State Estimation of Surrounding Vehicles Under Complex Connected Scenarios

Accurate and real-time acquisition of the motion state of surrounding vehicles is the premise of improving the handling stability and vehicle safety in dynamic complex scenarios. Whether the active safety technologies can run normally depends largely on the obtained vehicle states. Time-triggered mechanism is usually adopted to transmit sensor data between vehicles, which may lead to many issues such as information delay and data packet loss due to the limited communication bandwidth and energy constraints. To address the challenge, a dual event-driven cubature Kalman filter (DED-CKF) is designed to estimate the motion state of surrounding vehicles. A dual event-driven mechanism is introduced to adjust the data communication frequency. The communication transmission and packet loss trigger thresholds determine whether measurement information can be transmitted or not. The random variables satisfying the Bernoulli distribution are adopted to describe the occurrence of packet loss. Simulation and road tests show that the estimation precision of DED-CKF with 55.3% longitudinal acceleration communication rate and 75.9% lateral acceleration communication rate is 71.4% higher than CKF with complete data transmission under packet loss.