SMT Automatic Optical Inspection Path Planning Based on MDSPSO Algorithm

Abstract

When carrying out the Automatic Optical Inspection (AOI) on the Printing Circuit Board (PCB) adopted Surface Mounted Technology (SMT) , inspecting path planning is a question of combinatorial optimization, every inspecting window exists certain move range. In view of all above problems, this paper studied the standard Particle Swarm Optimization (PSO) algorithm, and based on that, it suggested another algorithm--Multi-directional Search PSO (MDSPSO) algorithm. The renew of the speed and location of every particle in MDSPSO algorithm, not only considered self individual extremal and overall situation extremal information, but also the information that other particles contained, which changed the unidirectional search in PSO algorithm into Multi-directional search, even improved the search accuracy. According to the convergence analysis, reaching the condition of ensuring the overall situation converges. Therefore the MDSPSO algorithm can converge fairly up to optimum result under the control of parameter in appropriate optional algorithm. By using the MDSPSO algorithm in SMT automatic optical inspection path planning, simulate result shows that, not only MDSPSO algorithm has merits of height accuracy and quick convergence speed, but also the quantity of inspecting windows of the path planning is the fewest in this algorithm and the inspecting route is the shortest.