This work presents a novel control
technique that combines linear
algebra-based controller (LABC) methodology with sliding surface concepts.
An LABC is developed from a first-order plus dead time model of the
process, which is improved to work under uncertainties by the use
of sliding surface concepts. Two different strategies are proposed
to reject constant and variable uncertainties. The result is two linear
controllers, which are tuned using four parameters at most. Results
of the control of two level tanks connected in series, a mixing tank
with variable dead time, and a laboratory batch reactor using this
novel technique are presented, including experimental and simulated
results. The efficiency of the proposed controller is tested under
nominal operating conditions and under parametric uncertainty and
persistent process disturbances. Proof of convergence to zero of tracking
errors is analyzed and included in this article.