Extension of the Application Potential of Wheeled Mobile Driving Simulators to Uneven Grounds

Driving simulators are an important element of vehicle development, since the design of driver assistance systems in particular requires the investigation of the driver-vehicle interaction. In the future, an even greater application potential is to be expected with regard to automated driving, since, for example, handover strategies can be investigated in a secure environment. However, today's driving simulator concepts have reached a limit with regard to the achievable quality of motion simulation. Especially urban driving scenarios require a range of motion that is not economically viable with the sled systems applied in current high-end systems. One way out of this limitation is provided by wheeled mobile driving simulators, which generate the demanded accelerations through tire forces. This enables an application on different driving surfaces, which allows flexible adaptation of the movement area to the requirements of the scenario. However, due to the contact between tire and driving surface, unevenness induces vibrations into the system which disturb the immersion of the subject. The known previous research on wheeled mobile driving simulators gathered in literature neglected this aspect and postulated a sufficient driving surface quality. However, it is unclear what sufficient means in this context. In addition, the flexibility advantage of the concept may be significantly limited by the requirement of a high quality surface. Thus, this work aims at quantifying the required driving surface quality and the development and evaluation of approaches for the reduction of disturbances induced by unevenness. First, an analysis of the current development state of the driving simulator at FZD, which includes a purely tire-sprung system with solid rubber tires, is conducted. This analysis shows that driving surface qualities with a maximum height deviation of 0.01 mm over a length of 4 m (so-called depth gauge) are required to use a driving simulator of this configuration without deteriorating the immersion of the subject. This quality is not achievable with asphalt surfaces, which offer the highest application potential for WMDS. The minimum achievable depth gauge amounts to 2 mm. Thereupon, an active compensation of the driving surface-induced vibrations with the Hexapod, which is already available in simulators, is investigated. The active approach increases the tolerable depth gauge by a factor of 4 compared to the passive tire-sprung system. Nevertheless, it is still only 3 % of the target value. Especially the high dead time of the hexapod as well as the low damping and the parameter fluctuations of the tire limit the potential of the concept. Therefore, the potential of implementing an additional suspension in combination with the active approach is investigated. In order to achieve a low natural frequency, which is advantageous in terms of vibration isolation, a kinematics is developed that reduces the suspension movements of the omnidirectional motion platform by support forces. In addition, the motion control of the driving simulator is adapted in order to adjust the wheel force distribution to the demands of the suspension. These measures reduce the disturbances caused by suspension movements to values below the perception threshold up to a horizontal acceleration of 4.5 m/s². The simulation of an urban driving scenario with a multibody model shows that this covers the majority of the occurring accelerations and that within more than 99% of the simulation time the disturbance motions remain below the perception threshold. With pneumatic tires, the acceleration range with ideal support can be increased to 5.4 m/s². With regard to the required driving surface quality, this allows an increase of the acceptable depth gauge to 0.8 mm, which corresponds to an improvement of almost two orders of magnitude compared to the initial situation. Nevertheless, the value is slightly below the minimum of 2 mm achievable with asphalt surfaces. However, the determined value is only required to remain below the perception threshold with the disturbance vibrations. As vibration in vehicles is not uncommon, the negative effects on the immersion could possibly be lower, allowing a slight exceeding of the threshold. Future subject studies must examine this aspect in more detail