Research On Vehicle Running Status Monitoring System

The goods are easily damaged during transportation,mainly due to vibration and impact.Large precision instruments,electronic products and other equipment sensitive to external forces are often damaged due to frequent impact in transit,causing losses to both buyers and sellers.Buyers,sellers and cargo carriers often dispute and controversy because of equipment damage.This thesis focuses on the vehicle status monitoring,mainly considering the indicators such as vehicle attitude angle and peak vertical acceleration.According to these indicators,the status of the vehicle can be judged when passing through bad road conditions such as potholes and bulges.At the same time,the monitoring data will be uploaded to the Ali Cloud platform for the receipt and delivery of goods and carrier three parties to view the data at any time,monitor the vehicle status and clarify the responsibilities of all parties.For this reason,the detection and evaluation of the impact in transit is particularly important.In this thesis,we focus on the problems of low accuracy of impact detection,poor real-time detection,and difficulties in the deployment of detection equipment,etc.The specific work is as follows.(1)The vehicle status monitoring system is designed.The MPU6050 inertial sensor and MT2503 D microprocessor are selected to build the data acquisition circuit.Through the SIM card module,the data is uploaded to the Ali Cloud platform and processed online to facilitate the monitoring of vehicle running status at any time.(2)The quaternion method is used to solve vehicle attitude for motion data.A statistical similarity measure cubature Kalman filter(SSMCKF)is proposed to solve the problem of low estimation accuracy caused by outliers in detection data.Based on the idea of correcting Kalman gain in SSMKF(Statistical Similarity Measure Kalman Filter),the problem of outlier sensitivity in Cubature Kalman Filter(CKF)is solved,and the correction factor is automatically optimized.Simulation results show that the proposed SSMCKF-Sqrt can adaptively reduce the interference of outliers,and has higher filtering performance and better numerical stability than CKF.(3)To achieve accurate tracking,a correct model needs to be established while the vehicle is running.The constant velocity model(CV),constant acceleration model(CA),Singer model and Jerk model are analyzed and compared.CA and Jerk target motion models are used.The problem that the estimation error of fixed Markov model transition probability is significant is analyzed.The model transition probability is modified in real time according to the ratio of model probability at the previous moment and the next moment,and a multi-model algorithm of adaptive model probability is proposed.The change of vehicle vertical acceleration under typical road conditions is analyzed and the identification experiment is completed.The experimental results show that the proposed algorithm can track the target in real time,effectively suppress the interference of general noise and outliers,and realize the accurate identification of road impact.