Compensation of Nonlinearities Using Neural Networks Implemented on Inexpensive Microcontrollers

This paper describes a method of linearizing the nonlinear characteristics of many sensors and devices using an embedded neural network. The neuron-by-neuron process was developed in assembly language to allow the fastest and shortest code on the embedded system. The embedded neural network also requires an accurate approximation for hyperbolic tangent to be used as the neuron activation function. The proposed method allows for complex neural networks with very powerful architectures to be embedded on an inexpensive 8-b microcontroller. This process was then demonstrated on several examples, including a robotic arm kinematics problem.