Intelligent PID Controller Based On FPGA

Artificial intelligence technology with the development of science and technology to improve, In recent years, the development direction of automatic control is intelligent control. As an evergreen in the control area, the leading role of the PID controller has become increasingly prominent.However, with industrial progress, the complexity of the controlled object and gradually increase.In the high requirement of control system such as time-varying, nonlinear, large delay, real-time and so on, traditional PID control method becomes powerless. Intelligent PID controller is in line with technological development, not only the algorithm is simple, to simplify the modeling procedure, but also has a self-learning, self-organizing, adaptive capacity, achieve efficient and reliable self-tuning PID controller. Therefore, the study of intelligent PID controller has far-reaching theoretical and practical value.In this paper, use artificial neural networks and intelligent optimization algorithms to optimized PID controller parameters, first careful analysis the two algorithms on the theoretical point of view,then use the platform of MATLAB optimization PID parameters, at last, to build intelligent PID controller by FPGA. Build FPGA hardware platform to avhieve intelligent PID controller.Radial Basis Function (RBF) neural network is an important member of the neural network family, is multi-layer static feedforward network, with a strong ability of function approximation and on-line identification, changes in the controlled object can be tracked in real time, and avoid fall into the local minimum value,very suitable for control and pattern recognition field.Using Radial Basis Function neural network to optimize traditional PID controller parameters, complementary advantages of both which can greatly improve the overall intelligence of the PID controller performance.Immune clonal selection algorithm is a adaptive artificial immune algorithm based on the combination of artificial immune and clonal selection, this algorithm has a strong self-learning and self-organizing capacity, can make the system with rapid response based on the given parameters and the basis of stability, is an intelligent control algorithm suitable for the complex object to optimize control parameters. Using immune clonal selection algorithm to optimize PID control parameters, can make PID controller's robust performance greatly improved. In this paper, first introduce RBF neural network and immune clonal selection algorithm, then using each of these two intelligent control methods to optimize the traditional PID control parameters and simulation by MATLAB.The simulation results shows that the RBF neural network and immune clonal selection algorithm of intelligent PID controller on the basis of taking into account the system dynamic and static performance, improve the accuracy and adaptive of PID control. Then build a hardware platform of the radial basis functionRBF neural network and immune clonal selection algorithm intelligent PID controller based on FPGA, the whole system calculation is more complex and computation, using only one module is not easy,so divided the large system modules into many small modules, and the small modules designed by the VHDL language or DSP Builder. After achieving design of each module, we construction closed-loop simulation based on DSP Builder and Simulink. And then carry out algorithm-level simulation in MATLAB2007, and RTL-level circuit simulation in Quartus II7.0and Modelsim SE6.2b, both algorithms run fast throughout the design process, full use of FPGA parallel computing and pipelining.The results show that the closed loop test of PID controller, this intelligent PID controller based on FPGA have flexible design, results show that the closed loop test of PID controller through DSP Builder, Simulink and Modelsim solves the sample source of test input as well as the input sample extraction of the controller, high reliability, self-tunning on line, low development cycle, robust and high speed, for the non-linear, time-varying and uncertainty complex industrial process,can obtain good control effect, therefore, the system control program feasible.