Design And Research Of Intelligent AGV For Custom Furniture Plate Transport

With the rapid development of Chinese economy,the customized furniture industry is growing explosively,and traditional furniture manufacturers are gradually being replaced by customized furniture manufacturers.Customized furniture is gradually becoming the mainstream of the market,and the demand is becoming higher and higher.In the large-scale production process of customized furniture manufacturers,the transportation efficiency of wooden boards directly affects the production efficiency of the manufacturers.At present,the transportation of wooden boards in the production process of customized furniture manufacturers is mainly carried out by humans.In the case of a large number of orders for products and a large weight of wooden boards,human transportation will consume a large amount of human resources,and it will also increase the probability of safety accidents in the production workshop.Aiming at the problems of low efficiency of wood transportation and great hidden danger of human transportation in the large order production process of customized furniture manufacturer an intelligent AGV for wood transportation of customized furniture is designed.The main design parameters of AGV are determined by analyzing the size,weight and maximum transportation number of wood.The guiding mode is laser guiding mode and the driving mode is two wheel differential driving mode,and the overall design scheme is determined.The design of various components,including chassis structure design and finite element analysis,board storage device structure design,drive and power system stress analysis,drive motor parameter calculation and selection,drive and power system structure design and other main components selection.The AGV dynamics is analyzed,including AGV straight action dynamics analysis and AGV turning dynamics analysis.The hardware and motion controller of AGV control system are designed,including:determining the overall scheme of the control system,single chip microcomputer and obstacle avoidance sensor selection,single chip microcomputer minimum working system circuit design,can bus communication interface circuit design,RS232 serial communication circuit design,RS485 serial communication circuit design,general I/O interface circuit design.The AGV kinematics model is established,and the fuzzy PID controller is designed as the motion controller.The path planning algorithm of AGV is studied.Q-learning algorithm of reinforcement learning algorithm is selected as the path planning algorithm.Aiming at the problems of slow convergence speed and many turns in the planning route of Q-learning algorithm,the original algorithm is improved by adding learning layer,adding prior knowledge of environment and adding turn penalty in reward function.The improved Q-learning algorithm is simulated.In order to verify the stability and feasibility of the algorithm,the simulation experiments are carried out in three different experimental environments.The experimental results show that the improved algorithm can quickly find a path with short path and less turns.Compared with Q-learning algorithm,adding learning layer Q-learning algorithm,introducing gravitational field Q-learning algorithm and deep double Q network algorithm,the improved algorithm has faster convergence speed and fewer turns.