Đang chuẩn bị liên kết để tải về tài liệu:
Robotics Automation and Control 2011 Part 14

Tham khảo tài liệu 'robotics automation and control 2011 part 14', kỹ thuật - công nghệ, cơ khí - chế tạo máy phục vụ nhu cầu học tập, nghiên cứu và làm việc hiệu quả | An Approach to Tune PID Fuzzy Logic Controllers Based on Reinforcement Learning 381 Factors PSe Level 0.75 selected PVSe PSde PVSde 0.12 0.3 0.17 TABLE VI. Modal points determined by discrete tuning The results of simulations fig.11 during the learning stage give an IAE criterion equals to 0.8544. We can notice then an improvement due to a larger space of solutions. Output voltage control Imposed Reference Current 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 _ . _ _ _ _ _ 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 time s Fig.11. tecedent tuning of the FPID controller Stage second consequent parameters tuning Initially to show the importance to tune the antecedents parameters we execute classic algorithm FQL with equally -distributed for antecedent memberships function. The results of simulations obtained during the learning stage give an IAE criterion equals 0.938 what corresponds to a deterioration of 10 of the optimization criterion as compared to the value obtained with classic FQL algorithm. Then we use the consequent FQL algorithm with the modal point values of the input membership functions determined in stage one.We get then the rule base showned in the table 7. The results of the simulations obtained during the learning fig 12 and 13 show an improvement of the IAE criterion equals now to 0.5318 compared to respectevely 0.8544 for the during the learning and validation phase. Therefore 382 Robotics Automation and Control we have reached an improvement of the criterion of 12 as compared to the stage one. These simulations results show an important decreasing of the optimisation criterion about 45 as compared to the value obtained before the tuning by FQL algorithms. More over the behaviour of tracking is improved because the response time is shorter. Output voltage control Imposed rference current Fig. 12. Consequent tuning of the FPID controller with classic FQL algorithm de NB NS NVS Z PVS PS PB e NB -1 -1 -1 -1 -0.2 -0.8 0 NS -1 -1 -1 -0.4 -0.4 0 0.8 NVS -1 -1 -0.8 .