The steering system is a critical component of all ground vehicles regardless of their propulsion source. Chassis directional control is provided by the steering system, which in turn relays valuable feedback about the road and vehicle behavior. As the primary feedback channel to the driver, the steering system also delivers the initial perception of a vehicle's handling and responsiveness to the consumer. Consequently, the steering system is an important aspect of the vehicle's evaluation and purchasing process, even if drivers are unaware of its direct influence in their decision making. With automobile purchases potentially hinging on the steering system, a need exists for a better understanding of steering preference through a focused research project. In this investigation, driver steering preferences have been studied using an advanced hardware-in-the-loop automobile steering simulator. Additionally, vehicle run-off-road situations have been studied, which occur when some of the vehicle wheels drift off the road surface and the driver recovers through steering commands.The Clemson University steering simulator underwent three significant generations of refinements to realize a state-of-the-art automotive engineering tool suitable for human subject testing. The first and third generation refinements focused on creating an immersive environment, while the second generation introduced the accurate reproduction of steering feel found in hydraulic systems and real-time adjustable steering feel. This laboratory simulator was the first known validated driving simulator developed for the sole purpose of supporting driver steering preference studies. The steering simulator successfully passed all validation tests (two pilot studies) leading to an extensive demographics-based driver preference study with 43 subjects. This study reflected the following preliminary trends: Drivers who used their vehicles for utility purposes preferred quicker steering ratios and heavier efforts in residential, country, and highway environments. In contrast, car enthusiasts preferred quick steering ratios in residential and country environments and light steering effort on the highway. Finally, rural drivers preferred quicker steering ratios on country roads. These relationships may be used to set steering targets for future vehicle developments to accurately match vehicles to their intended market segments.The second research aspect was the development of an objective steering metric to evaluate a driver's steering preference. In past simulator studies, driver feedback has been gathered extensively using written questionnaires. However, this delays the testing procedure and introduces an outside influence that may skew results. Through the data collected in this project, a robust objective steering preference metric has been proposed to gather steering preferences without directly communicating with the driver. The weighted steering preference metric demonstrated an excellent correlation with survey responses of r=-0.39 regardless of steering setting. This global steering preference metric used a combination of yaw rate, psi, longitudinal acceleration, ax, and lateral acceleration, ay. The objective data was further dissected and it was discovered that changes made to the steering ratio resulted in a correlation of r=-0.55 between the objective data and subjective response from the test subjects. This substantial correlation relied on the longitudinal acceleration, ax, left front tire angle, deltalf, and throttle position, TPS.Beyond steering preferences, vehicle safety remains a major concern for automotive manufacturers. One important type of crash results from the vehicle leaving the road surface and then returning abruptly due to large steering wheel inputs: road runoff and return. A subset of run-off-road crashes that involves a steep hard shoulder has been labeled "shoulder induced accidents". An active steering controller was developed to mitigate these "shoulder induced accidents". A cornering stiffness estimation technique, using a Kalman filter, was coupled with a full state feedback controller and "driver intention" module to create a safe solution without excessive intervention. The concept was designed to not only work for shoulder induced accidents, but also for similar road surface fluctuations like patched ice. The vehicle crossed the centerline after 1.0s in the baseline case the controller was able to improve this to 1.3s for a 30% improvement regardless of driver expertise level. For the case of an attentive driver, the final heading angle of the vehicle was reduced by 47% from 0.48 rad to 0.255 rad.These laboratory investigations have clearly demonstrated that advancements in driver preference and vehicle safety may be realized using simulator technology. The opportunity to apply these tools should result in better vehicles and greater safety of driver and occupants. With the development of the objective steering preference metric, future research opportunities exist. For prior steering preference research, the feedback loop has typically required interaction with the subject to rate a setting before continuing. However, the objective steering preference metric allows this step to be automated, opening the door for the development of an automatic tuning steering system.
展开▼
