FPGA implementation of a recurrent neural fuzzy network for on-line temperature control

FPGA implementation of a TSK-type recurrent neural fuzzy network (TRNFN) for water bath temperature control is proposed in this paper. The TRNFN is constructed from recurrent fuzzy if-then rules and is built through a concurrent structure and parameter learning. To apply TRNFN to temperature control, the direct inverse control configuration is adopted. For the on-line adaptive control objective, the implemented TRNFN chip is characterized with learning ability, where the consequent part parameters of TRNFN are tuned by gradient descent. Experiments on water bath temperature control have verified the function of the designed TRNFN chip.