The control of electric motors using intelligent techniques is important in improving the motor's efficiency, to make alternative fuel vehicles a realistic consumer option. Until recently, electric motors used in automotive applications were not designed to operate at their peak efficiency at various output torques, and mechanical transmissions can consume a lot of the energy. Recent improvements in motor technology would permit motors to function at or near peak efficiency for a wide variety of parameters. However, the parameters are varying constantly, so that only an adaptive control scheme would be able to achieve such efficiency. ORINCON proposes to develop an artificial neural network (ANN) model reference control scheme that will adapt the motor to behave in the most efficient manner possible given the current conditions. Unlike most ANNs, this neuro-controller will be trained on-line using an extended Kalman filter trainer that has shown great potential.
