A novel 3D (error, derror, dderror) neurofuzzy control architecture

The paper presents a novel high order derivative fuzzy control architecture. This architecture uses a three input error, derror, dderror structure which has proven to be successful in controlling a double inverted pendulum system. The control system may be automatically generated from the required granularity of the input and output variables, making the specification of the control rules a less tedious process. A neural network tuning process is applied to the fuzzy controller in order to optimise the performance. The novel three dimensional (error, derror, dderror) fuzzy controller application to the double inverted pendulum in real time on a physical system is presented in detail.