Hardware/Software co-design of an ADALINE based adaptive controller for a DC motor

We present here results pertaining to an adaptive linear neuron (ADALINE) based PID control system co-implemented in a software/hardware approach. The presented method primarily designs a PED controller in the discrete time domain with a simplified mathematical model. The parameters of the system model are online estimated using an ADALINE neural network. Due to online estimating the system parameters and momentary tuning the PID gains, a fast data processing approach is needed to implement the design. In our scheme, an FPGA chip is used to obtain and process the measured data. Further, a microprocessor is explored to implement core computation for the algorithm. The designed system is tested using a permanent magnetic DC motor. Experiment results verify the effectiveness of the proposed methodology. Moreover, the setup of the experimental bed indicates that the adaptive control system can be tested using low-cost devices with satisfactory performance due to mutual benefit of software and hardware in the co-design approach.