[Paper Review] A Direct Slip Ratio Estimation Method based on an Intelligent Tire and Machine Learning

Nan Xu, Zepeng Tang+3 more|arXiv (Cornell University)|Jun 9, 2021

Vehicle Dynamics and Control Systems4 citations

TL;DR

This paper proposes a novel direct slip ratio estimation method using an intelligent tire equipped with a triaxial accelerometer and machine learning algorithms. By leveraging vertical acceleration as a key feature, the method achieves highly accurate, continuous slip ratio estimation across varying vehicle speeds and loads, with the best model (ANN) achieving a normalized root mean square error (NRMS) of 4.88% via 10-fold cross-validation.

ABSTRACT

Accurate estimation of the tire slip ratio is critical for vehicle safety, as it is necessary for vehicle control purposes. In this paper, an intelligent tire system is presented to develop a novel slip ratio estimation model using machine learning algorithms. The accelerations, generated by a triaxial accelerometer installed onto the inner liner of the tire, are varied when the tire rotates to update the contact patch. Meanwhile, the slip ratio reference value can be measured by the MTS Flat-Trac tire test platform. Then, by analyzing the variation between the accelerations and slip ratio, highly useful features are discovered, which are especially promising for assessing vertical acceleration. For these features, machine learning (ML) algorithms are trained to build the slip ratio estimation model, in which the ML algorithms include artificial neural networks (ANNs), gradient boosting machines (GBMs), random forests (RFs), and support vector machines (SVMs). Finally, the estimated NRMS errors are evaluated using 10-fold cross-validation (CV). The proposed estimation model is able to estimate the slip ratio continuously and stably using only the acceleration from the intelligent tire system, and the estimated slip ratio range can reach 30%. The estimation results have high robustness to vehicle velocity and load, where the best NRMS errors can reach 4.88%. In summary, the present study with the fusion of an intelligent tire system and machine learning paves the way for the accurate estimation of the tire slip ratio under different driving conditions, which create new opportunities for autonomous vehicles, intelligent tires, and tire slip ratio estimation.

Motivation & Objective

• To develop a direct, real-time slip ratio estimation method that does not rely on vehicle velocity sensors or complex kinematic models.
• To leverage an intelligent tire system with embedded accelerometers to extract dynamic features for slip ratio estimation.
• To evaluate the performance of multiple machine learning algorithms in estimating slip ratio under diverse driving conditions.
• To identify the most effective input features—particularly vertical acceleration—for high-accuracy slip ratio estimation.
• To validate the robustness and generalization capability of the model using 10-fold cross-validation across multiple datasets.

Proposed method

• An intelligent tire system was developed with a triaxial accelerometer mounted on the inner liner to capture real-time acceleration signals during tire rotation.
• The MTS Flat-Trac tire test platform was used to measure ground truth slip ratio values under controlled conditions.
• Key features were extracted from the acceleration signals, with vertical acceleration (az) identified as the most informative for slip ratio estimation.
• Four machine learning models—Artificial Neural Networks (ANN), Gradient Boosting Machines (GBM), Random Forests (RF), and Support Vector Machines (SVM)—were trained using the extracted features.
• 10-fold cross-validation was applied to evaluate model generalization and robustness across different vehicle speeds and loads.
• Model performance was quantified using normalized root mean square error (NRMS), with ablation studies comparing different input combinations (e.g., (ax, az), az alone).

Experimental results

Research questions

• RQ1 Can an intelligent tire system with embedded accelerometers provide sufficient dynamic features for direct slip ratio estimation without relying on vehicle velocity sensors?
• RQ2 Which machine learning algorithm—ANN, GBM, RF, or SVM—delivers the most accurate and stable slip ratio estimation under varying driving conditions?
• RQ3 How does the inclusion of specific acceleration components (e.g., vertical az) affect estimation accuracy and computational efficiency?
• RQ4 What is the robustness of the slip ratio estimation model across different vehicle speeds, loads, and slip ratio ranges (up to 30%)?
• RQ5 Can a simplified input set (e.g., only az) maintain high accuracy while reducing computational cost for real-time implementation?

Key findings

• The proposed method enables continuous and stable slip ratio estimation using only accelerometer data from an intelligent tire, with no reliance on vehicle velocity sensors.
• The artificial neural network (ANN) achieved the lowest normalized root mean square error (NRMS) of 4.88% when using vertical acceleration (az) as input, demonstrating the highest accuracy.
• The gradient boosting machine (GBM) model showed comparable accuracy with the smallest variance in NRMS error, indicating high robustness and stability.
• Vertical acceleration (az) alone provided superior estimation performance compared to combined inputs (ax, ay, az), reducing computational load while maintaining high accuracy.
• The model demonstrated strong robustness across varying vehicle speeds and loads, with consistent performance even at high slip ratios up to 30%.
• The 10-fold cross-validation confirmed that the model generalizes well to unseen data, with the (az) input configuration yielding the most favorable trade-off between accuracy and computational efficiency.