Abstract This article aims at showing how neural networks can be employed in the generation of human-machine interfaces (neurointerfaces) for practical real time applications. In a great number of real world applications, due to technical and economic factors, full automation is not possible. In such cases, the human presence is essential and indeed, the system performance becomes highly dependent on human skills. Accordingly, an interface that modifies the problem, allowing unskilled human operators to perform the same task in a satisfactory way, becomes extremely useful. The adaptive nonlinear inverse modeling approach is employed as the basic methodology for specification and design of neurointerfaces. A successful application, a neurointerface that helps an operator to back up a scaled truck model connected to single-trailer and double-trailer configurations, is presented.
[1]
P. Werbos,et al.
Beyond Regression : "New Tools for Prediction and Analysis in the Behavioral Sciences
,
1974
.
[2]
Geoffrey E. Hinton,et al.
Learning internal representations by error propagation
,
1986
.
[3]
Derrick H. Nguyen,et al.
Neural networks for self-learning control systems
,
1990
.
[4]
M. M. Lamego,et al.
The interval based control technique: controlling physical systems through imprecise models
,
1995,
IAS '95. Conference Record of the 1995 IEEE Industry Applications Conference Thirtieth IAS Annual Meeting.
[5]
Paul J. Werbos,et al.
Backpropagation Through Time: What It Does and How to Do It
,
1990,
Proc. IEEE.
[6]
Adam Farquhar,et al.
Modeling and Simulation of a Complex Industrial Process
,
1997,
IEEE Expert.