A novel maximum correntropy extended Kalman filter with event-triggered mechanism for state estimation of preceding vehicles in V2V environments

This study addresses the state estimation problem for the preceding vehicle in vehicle-to-vehicle (V2V) cooperative perception, with the objective of balancing estimation accuracy against communication load under constraints of limited bandwidth and non-Gaussian noise interference. To this end, an extended Kalman filtering algorithm is proposed that integrates an event-triggered (ET) mechanism with the maximum correntropy criterion (MCC). The event-triggered mechanism adaptively regulates communication frequency, significantly reducing network load while preserving estimation performance. Meanwhile, the maximum correntropy criterion enhances the filter’s robustness against non-Gaussian noise, enabling high-precision state estimation under limited communication resources. Simulation results under double lane-change and continuous sinusoidal steering scenarios show that, compared to conventional extended Kalman filtering and existing event-triggered robust filtering methods, the proposed algorithm achieves better estimation accuracy and stability in non-Gaussian noise environments. Moreover, it reduces the average communication load by approximately 68% without compromising estimation performance. This work provides an effective solution for state estimation in V2V cooperative perception that combines high robustness with high communication efficiency.