| Motion control has become a research hotspot in the field of intelligent manufacturing.Breakthrough in key issues of motion control are important basis for intelligence machine.Focusing research on these issues will help to solve common problems and enhance the competitiveness of our country on this field.They are also difficult problems for integrating the knowledge of mechatronics,computer,automatic control and related fields.In recent years,new ideas and methods have been proposed to solve them,through the integration of computational intelligence,big data and microelectronics technology.This thesis will mainly research on the topics of velocity control,trajectory planning and indoor localization in motion control.Acceleration and deceleration algorithms are the core issue of velocity control.A key problem in high-speed and high-precision intelligent equipment is how to reduce the resource consumption of flexible control algorithm as well as improving the control accuracy and real-time performance under the constraint of hardware resource.Issues related to trajectory planning issues are as important as velocity control issues and so,they will be considered synchronously in motion control research.How to optimize the constraint of multiple factors is a valuable point of these research.Location is also a basic information of motion control in indoor environment.A challenging issue in this research is how to reduce the non-linear impact of a localization model while enhancing the robustness and improving the location accuracy.Below are the main contributions of this thesis:(1)This thesis has proposed an iterative velocity control algorithm,called Tri_Iteration,which has an authorized a patent of invention(ZL201510955082.X).Firstly,the common acceleration and deceleration algorithms are analyzed and compared regarding aspect of efficiency on acceleration,flexible control and calculation.Then,the mainly work studied a flexible velocity control algorithm based on trigonometric function.Aiming at the constraint of hardware resource for trigonometric function calculation in FPGA chips,a velocity control model is proposed,which transforms trigonometric function calculations into addition,subtraction and multiplication operations by an iterative method.Suchmethod avoids the storage of large amount of trigonometric function values.It saves memory resource and helps to improve the control accuracy.It also does not require direct calculation of the value of trigonometric function,which reduces the running clock cycle and improves the real-time performance of the algorithm.Finally,the proposed model was implemented and deployed on a FPGA chip,resulting in an iterative circuit.After initial parameters are inputted,it can generate the velocity control values continuously.The experimental results show that it has a significant effect in improving the control accuracy and saving hardware resources.It reveals as a helpful research work to improve the velocity control accuracy and real-time performance for intelligent device.(2)This thesis has proposed a planning method of trajectory and velocity based on PSO algorithm,which is called DRAMATIC(Dynamic Trajectory and Velocity Planning Method).Firstly,the feasibility of PSO algorithm utilized in trajectory and velocity planning was analyzed.Then,main works were focused on the constraint of multiple factors between the control accuracy and kinetics.Next,a model based on PSO algorithm was designed to solve this issue.Chord error was applied in the fitness function to select the particle.After calculating all the satisfied points,the continuous line segments will be constructed.According to the Trigonometric Acceleration and Deceleration algorithm,the kinetic parameters were estimated,including velocity,acceleration and jerk.The experimental results show that it has a significant effect in saving time and optimizing the constraint of multiple factors.It reveals as a beneficial research work to improve the processing accuracy and efficiency for intelligent device.(3)This thesis has proposed an indoor localization method based on ANN(Artificial Neural Network),called BP_Landmarc(Indoor Localization Algorithm Based on BP and Landmarc),which has been applied a patent of invention.Firstly,the indoor localization methods based on RSSI(Received Signal Strength Indication)and RFID(Radio Frequency Identification)technologies are analyzed and compared.Then,putting emphasis on the issue of accuracy caused by non-linear impact in localization algorithm,a model based on BP ANN was constructed.Reading the RSSI values of the reference tags,the model takes them as input data alongside the actual coordinates of reference tags as the supervised data,to train the BP ANN.The trained model can calculate the coordinate of the tracking RFID tags.Furthermore,aiming to increase the robustness,different indoor environment factors were taken as model's working parameters.An indoor localization algorithm was realized based on this works.The experimental results show that it can reduce the negative influence of non-linear factors,improving the localization accuracy.Main works related to the above three issues are successfully and fully implemented and tested.These works greatly improve the calculationefficiency by utilizing the relativity of sequence velocity values,while providing a novel method to realize high velocity and accurate flexible control under constrained hardware resource.These works explore new approach to enhance real-time performance and dynamic adaptability for intelligent device through the integration of knowledge on computational intelligence,machine learning and Intelligent Manufacturing and so,breaking the technical bottleneck of existing models and methods.Integrating the above research results,a basic motion control system can be constructed,which can be beneficial for problems and production efficiency in intelligent manufacturing.Simultaneously,it has great significance to enhance the independent intellectual property rights of key technologies in intelligent manufacturing field. |
PDF Full Text Request