The hydraulic clutch actuation path used in heavy duty transmissions often shows a lot of variability due to manufacturing tolerances and ageing effects. Reason for this are in particular varying friction coefficients in the spools and external factors such as compliance with the specified service intervals or the choice of hydraulic fluid. As a direct consequence, the shift quality typically varies from one transmission to the next. To resolve this problem, this paper presents a machine learning algorithm for the feedforward control of the hydraulic clutch actuation path, a model-based and a data-based feedforward approach. The two approaches are evaluated and compared with each other in simulations with a high-fidelity model. As it turns out, the model-based version is to be preferred and therefore used in a real world evaluation on an embedded controller and a 16 tons wheel loader.

Hydraulic clutch actuation paths in heavy duty transmissions show a lot of variability due to series production variance, e.g. different friction. Additionally, there are external factors, such as the operators choice of hydraulic fluid or the compliance with specified service intervals. A further factor is the amount of entrained air that can vary from clutch to clutch in a single transmission. To address these effects, the authors recently proposed feedforward control concepts for the hydraulic clutch actuation path. While the feedforward controllers were shown to significantly improve the control quality, a manual parameter adaptation is necessary for optimal performance. Therefore, this paper presents and evaluates different learning algorithms. One algorithm based on rules and one based on multi-armed bandits that are suitable for a model-based control scheme. The third, a reinforcement learning algorithm based on policy gradients is suitable for a data-based feedforward control concept. In simulations with a high fidelity model, the improvement achievable in comparison with the state-of-the-art approach with all three algorithms is shown. In addition, the strengths and weaknesses of each algorithm are discussed.

A nonlinear MPC framework is presented that is suitable for dynamical systems with sampling times in the (sub)millisecond range and that allows for an efficient implementation on embedded hardware. The algorithm is based on an augmented Lagrangian formulation with a tailored gradient method for the inner minimization problem. The algorithm is implemented in the software framework GRAMPC and is a fundamental revision of an earlier version. Detailed performance results are presented for a test set of benchmark problems and in comparison to other nonlinear MPC packages. In addition, runtime results and memory requirements for GRAMPC on ECU level demonstrate its applicability on embedded hardware.

This paper presents a nonlinear model predictive control (MPC) strategy for heavy-duty multigroup transmissions. The underlying optimal control problem consists of three phases with unspecified end times to deal with the state-dependent switching of the hybrid dynamics and additionally accounts for several state and input constraints. The design objective is to maintain the output speed and output torque while minimizing the time needed to shift gears. A tailored nonlinear optimization algorithm on a shrinking horizon is presented in order to achieve real-time feasibility of the MPC scheme on an electronic control unit (ECU). Simulations against a high-fidelity model and runtime results demonstrate the performance of the MPC scheme for gear downshift and upshift as well as the algorithmic efficiency with computation times in the single-digit millisecond range on an ECU level. The real-world applicability is demonstrated by presenting experimental results for a 16 tons wheel loader.

The hydraulics of heavy-duty transmissions typically lack pressure sensors, which implies that the hydraulic actuation system has to be controlled in a feedforward manner. In practical applications, this is often done by numerically inverting a static mapping, leading to undesired pressure errors and inaccurate clutch positioning especially in dynamic situations. This paper therefore investigates two feedforward control strategies for a hydraulic clutch actuation path. The first is a flatness-based feedforward control strategy that relies on a high accuracy model, and additionally accounts for input constraints. The second one is a data-driven neural network-based feedforward control strategy that relies on measurement data. The feedforward controls are first evaluated in simulations and subsequently evaluated experimentally for a real-world transmission that demonstrates the realizability on ECU level as well as the effectiveness of the flatness-based and neural network-based feedforward controls compared to the standard static mapping-based approach.

This paper derives an accurate dynamical model of an indirectly controlled hydraulic actuation path for a wet friction clutch. The considered setup exhibits a relatively low pressure range, which necessitates to account for the variable compressibility of the hydraulic fluid due to entrained air. In addition, the model design accounts for the complex valve geometry as well as the draft angles of the valve housing due to injection moulding. The relevant hysteresis phenomena of the valves are modelled via Prandtl-Ishlinskii and LuGre operators, respectively. Simulation results show the accuracy of the model with an average error of 0.9% (180 mbar) compared to measurements. In addition, the necessity to respect the entrained air is verified by a counter-simulation with constant bulk modulus.

The paper presents an optimal control strategy for a heavy-duty multi-group transmission that allows for simultaneous shifting in all groups. The approach is presented for a two-group transmission without loss of generality. The optimal control problem (OCP) consists of three phases with state-dependent switching of the hybrid dynamics and free end-time characteristics as well as several input and state constraints. The objective is designed to minimize the gear shift time while maintaining the speed and torque at the output shaft. A tailored optimization algorithm is presented to numerically solve the nonlinear OCP in a highly efficient manner. A comparison with a standard SQP method as well as simulation and runtime results on CPU and ECU level demonstrate the performance of the algorithm for an embedded real time implementation.