Study On Static Decoupling And Signal Denoising For Multi-axial Wheel Force Transducer

The movement of automobile is caused by the mutual force effect between the ground and the wheels, and it is of great significance to measure the multi-axis wheel forces for researching automobile and its subsystems. The wheel forces are measured using a muti-axial wheel force transducer (WFT) as the key technique, and it is valuable in the theory research and engineering applications to improve the measurement precision of WFT. The paper studies static decoupling and signal denoising with the purpose of improving the measurement precision of WFT. The main research work is summarized as follows:(1) The static calibration tests are finished using the hydraulic calibration device for WFT. According to the characteristics of the calibration data, the paper proposes the method using the wavelet transform to eliminate gross errors and random errors synchronously. Then filtering preprocessing is performed to the calibration data using the proposed method, and good filtering results are achieved.(2) Assuming WFT is a linear system, the paper does significance tests to verify WFT system's linear hypothesis. Considering single axis force cannot be put on WFT indeed during the calibration tests, the paper puts forward iterative method to realize linear decoupling, and then characteristics of crosscoupling are analyzed. Data obtained from actual driving conditions in the automobile road experiment is linearly decoupled, and the experimental results show that linear decoupling for WFT eliminates linear crosscoupling error and improves the measurement precision.(3) To overcome the limitations of static linear decoupling, static nonlinear decoupling of WFT is studied. The paper proposes a static nonlinear decoupling method based on GA-BP artificial neural network. The emulation tests are done using the sample calibration data, and the emulation results indicate that the method is feasible and effective. The proposed method develops ideas of decoupling for WFT.(4) Having analyzed the signal-to-noise characteristics of WFT, the paper adopts the wavelet transform as the denoising method. The conventional threshold denoising method and the ameliorated threshold denoising method based on the 3σrule are used in denoising WFT signals respectively. The experimental results show that obvious effect can be obtained using threshold denosing method for WFT signals, and the improved threshold method outperforms the conventional threshold method for WFT signals denoising under the same wavelet function and decomposed wavelet levels.