New Road Hazard Classification Enabled by Rack Force Estimation of Electric Power Steering Systems

Abstract

This study presents a preliminary investigation of a new strategy for detecting and classifying road hazards, such as potholes and bumps, based on rack force estimation of electric power steering (EPS) systems. The numerous studies on road hazard detection have primarily focused on computer-vision systems, including cameras or light detection, and ranging and vertical vibration signals measured by accelerometers mounted on suspension systems. However, conventional methods are prone to reduced accuracy owing to their susceptibility to vibrations transmitted from road surfaces to vehicles. Accordingly, considerable room for detection accuracy improvement remains. Herein, we explore a novel approach that leverages the steering rack force generated in the EPS system, considering that potholes and bumps induce vertical and lateral forces on the tire’s contact patch, resulting in a net force generated by such tire moments on the vehicle steering rack. We propose an algorithm that uses an improved Kalman filter (KF) with an unknown input, combining the capabilities of a conventional KF with a disturbance observer. This algorithm aims to estimate the rack force by utilizing measurements of the steering torque and angle inputs. The estimated rack force provides features that serve as the basis for classifying road hazards. The classification performance was evaluated using metrics calculated from confusion matrix, such as accuracy, precision, recall, and F1. The proposed road hazard detection and classification algorithm was not only rigorously simulated using SIMULINK® with CarSim® software, but also is experimentally validated through in-vehicle tests.

Keywords

• Rack force
• electric power steering system
• road hazard
• Kalman filter with unknown input
• support vector machine
• in-vehicle test