Hydraulic cylinder drives with coupled control edges
suffer from inherent losses due to dependent meter-in and
meter-out flow control. Velocity and pressure demands with
different valve opening requirements collide and result in energy
dissipation at the control edges. Independent metering
technology provides individual actuation of inlet and outlet
valves and leads to higher efficiency. The additional control
input offers separate control of cylinder velocity, i.e. flow, and
chamber pressure. However, the resulting nonlinear Multiple
Input Multiple Output (MIMO) system poses several challenges
for the control algorithm. Cross coupling between the system
states requires multivariable control. Furthermore, the commonly
used linearization approaches strongly depend on the current
operating point of the hydraulic system. In mobile application,
these operating points are usually difficult to predict and move
across a wide range. Therefore, a nonlinear, multivariable control
algorithm is required. This paper presents some frequently
discussed nonlinear control approaches, namely flatness-based
tracking control, sliding mode control and exact input/output
linearization, and evaluates their performance and applicability
in the mobile hydraulic context. Control variables are velocity
and weighted chamber pressure of the cylinder. Previous research
on these controller concepts showed promising results for simple
nonlinear MIMO systems. However, more complex systems,
incorporating modelling uncertainties and limited observability
have rarely been studied thoroughly. Therefore, this study focuses
on the validation of the controller implementation based on
simplified models of a hydraulic valve-controlled manipulator
with two joints. This is followed by an investigation of the
performance on increasingly sophisticated and realistic models.
Assessment criteria are robustness against parameter and model
uncertainties, computational efficiency, communication delay and
measurement effort.
Index Terms—Nonlinear Control Theory, Flatness-based Control,
Sliding Mode Control, Exact Input/Output Linearization,
Independent Metering, Hydraulic Manipulator