FCMAC: a fuzzified cerebellar model articulation controller with self-organizing capacity
Automatica (Journal of IFAC)
A fuzzy CMAC model for color reproduction
Fuzzy Sets and Systems
Temporal credit assignment in reinforcement learning
Temporal credit assignment in reinforcement learning
Stability and weight smoothing in cmac neural networks
Stability and weight smoothing in cmac neural networks
A comparison of natural and artificial intelligence
ACM SIGART Bulletin
Embedding fuzzy mechanisms and knowledge in box-type reinforcement learning controllers
IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics
Hi-index | 0.00 |
This paper presents an online neural network controller. Cerebellar Model Articulation Controller (CMAC) is suitable to online control due to its fast learning speed. By integrating the CMAC address scheme with fuzzy logic concept, a general fuzzified CMAC (GFAC) is proposed. Then by incorporating the concept of eligibility into the GFAC, a GFAC controller with eligibility is presented, named FACE. A learning algorithm for the FACE is derived to tune the model parameters. To achieve online control, an efficient implementation of the proposed FACE method is presented. As an example, the proposed FACE is applied to a ship steering control system. The simulation results show that the ship course can be properly controlled under the disturbances of wave, wind and current.