This work represents an innovative approach for the control of a Limited Slip Differential (LSD). The limited slip differential transmits the power of the motor to the ground allowing the wheels to spin at different speed. Its task is dividing the transmitted torque between the driven wheels in different driving situations and scenarios. We start with covering the current trends and an introduction to the functionality and impact of a limited slip differential in driving dynamics. Since the current control system for such a differential is very complex and has no ability to adapt itself over time to the changes, this work proposes a new control approach, based on machine learning techniques. Due to the features of the data sets, gathered from real driving situations and are used for the training of the model, the supervised regression-based machine learning methods are selected for evaluation. To be able to choose the right regression method, the data for training the model is closely analyzed and an appropriate model that has the ability of improving the accuracy of a limited slip differential control while ensuring a safe, pleasant and high performance drive is chosen.


    Access

    Download


    Export, share and cite



    Title :

    A Regression-based Control Approach for Limited Slip Differencial


    Contributors:

    Publication date :

    2017-12-31


    Type of media :

    Paper


    Type of material :

    Electronic Resource


    Language :

    English


    Keywords :

    Classification :

    DDC:    629 / 000



    Ship manoeuvrability analysis using the differencial approach

    Asinovsky, V. / Landsburg, A.C. | Tema Archive | 1987


    Fault Detection of Nonlinear Process Based on Differencial Locality Preserving Projections

    Guo, Jinyu / Zhong, Lulu / Li, Yuan | British Library Online Contents | 2018


    Limited slip braking

    McLoughlin, J.H. | Tema Archive | 1985


    Limited Slip Differentials

    Frömmig, Lars | Springer Verlag | 2023


    Limited slip differential

    FOX MATTHEW G / HEATWOLE GREGORY L / EDLER ANDREW N et al. | European Patent Office | 2015

    Free access