TU Berlin

Chair of Electronic Measurement and
Diagnostic Technology
Nonlinear control strategies in automotive powertrains

Page Content

to Navigation

Nonlinear control strategies in automotive powertrains

The available ressources and performance of automotive control units is increasing. So the following question arises

"Can the performance and the system behaviour of modern transmissions be increased by the use of modern nonlinear control strategies?"

and is the base of this research topic.

In the first step the transmission with its (double) clutch or torque converter is modeled and a analytic controller design is done
On this way the apropriate controller strategies will be selected which will be implemented on a rapid-prototyping-control-unit. This phase is followed by different test cycles on the transmission test bench. So near-to-reality measurements results will be obtained. The most promising strategies will be test on a vehicle.

Stability behaviour, control quality and disturbance suppression are the characteristics to evaluate the nonlinear control strategies.

Possible nonlinear control strategies

For the clutch and gear shifting control the following strategies should be implemented:

• Flatness based approach

Differentially flat systems have a number of interesting properties which make a control of nonlinear flat systems very easy. When a flat output for a nonlinear system is found a classical control design is possible. In addition to that the motion planning for the flat output is very easy and system constraints can easily integrated.

• Sliding-mode-control

In sliding-mode-control different nonlinear feedback paths are integrated into the closed loop structure and state-depended switchovers between these paths is done. The so called switching surface S(x) is the base for the switch decision. One characteristic property of sliding-mode-control is the direction of the system trajectories: All trajectories point towards the switching surface. When infinitely quick switching is supported by the actuators the system moves on this switching surface into the steady-state once S(x) is reached.
Unfortunately indefinitely quick switching is impossible in reality. This is the reason for chattering-phenomena that stresses the actuators. Perhaps is chattering is an advantage for the control of electro-hydraulic actuators because this might cause a softer transition from stick to slip behaviour.

• Gain-scheduling

Since the 1960's gain-scheduling is successfully used in control of nonlinear system in aeroplane and rocket applications. In gain-scheduling-approach the system state space is divided into small regions where the system behaviour could be approximated with linear functions. For each linear approximation a classical controller design is done. In contrast to structure variable and sliding-mode-approaches the transitions between different controller structures isn't done by a hard switchover. Instead continuous functions for the controller parameters are interpolated which depend on the scheduling states. So a smooth transition between the different controllers is achieved.


Quick Access

Schnellnavigation zur Seite über Nummerneingabe