Non-iterative implicit integration methods with linearization-based equations of motion approximation for real-time multibody vehicle dynamics simulation

The growing importance of virtual vehicle development and subjective assessments of driving characteristics on driving simulators pushes the demand for real-time vehicle dynamics simulation. Real-time capable high-fidelity multibody simulation models are required that capture the influence of the kinematics and elastokinematics of wheel suspensions on driving dynamics accurately. Elastic bushings, commonly used in wheel suspensions to enhance ride comfort and handling, introduce high numerical stiffness of the equations of motion and pose significant challenges for real-time simulation. Non-iterative implicit integration methods offer an attractive solution by combining stability, accuracy and deterministic computational effort. These methods avoid iterative solutions of nonlinear equation systems by exploiting the Jacobians of the equations of motion, making them well suited for real-time applications. To reduce their computational effort, this study simplifies several non-iterative implicit integration methods with a linearization-based equations of motion approximation that is updated every few integration steps. The performance of the methods is evaluated and compared for real-time multibody simulations of a double wishbone suspension with elastic bushings. The influence of the integration step size and the update frequency of the equations of motion evaluation and linearization on computation times and accuracy is analyzed using dynamic suspension tests.