Programming a Six Axis Motoman Hp3c Robot for Indus- Trial Sorting Application

Industrial systems efficiency often can be improved by replacing human subjects with robotics technology. With the virtue of automation couple with robotics, this picture is gradually being visible. Accurate inspection of a product in its all stages of production has become a very crucial part of the packaging industry. Nearly 100% accurate inspection and detection can be achieved by using high speed camera vision systems coupled with robotic arms. However the investmentto-yield ratio must make sense. In this research the proposed setup of camera vision system along with robotic arm will inspect multiple patterns of defects with slight modification in the inspection system. The impetus of this research originated from a common problem encountered in the manufacturing and packaging industry. Generally, an inspecting conveyor system requires a human operator to visually inspect the part or in case of an automatic detection, the process is limited in its inspecting ability. In case of human operator inspecting process, the operator requires to be trained in the particular inspection criteria as well as the use of precision handtools to determine the defect. This involves training cost, overhead expenditure (higher pay-scale due to training and experience), and other labor fees for a monotonous, repetitive and no-brainer job. In case of existing automatic systems, the detection method is limited and not adoptable to various types of parts and their defects. For packaging industry and its products, quality control is one of the most important steps of manufacturing process. The objective of this research project is to build an automated inspection setup consisting of a high speed camera vision system, an industrial robot, a servomotor controlled conveyor system, and a pneumatic gripper. The control setup will inspect and sort parts based on its acceptability using a high speed machine vision system. In this research project, the inspection system consists of a Motoman (produced by Yaskawa) 6-axis articulated industrial robot, a Cognex [5] high speed camera vision system and a servomotor driven conveyor are replacing an existing error-prone inspection stage that will improve the degree of accuracy by three fold. The proposed inspection system is intelligent enough to recycle parts that are marginally acceptable and worthwhile to be modified as good parts. Introduction: In packaging industrial, sorting, inspections and separations are one of the most extensively researched subject matters. Deflect detection and product separations are considered to be the top priorities in any manufacturing environment. For speeding up sorting operation robots are used extensively in industry. Many industrial sorting applications robotic technology adds special benefits to the process. When it comes to improved level of inspection accuracy and speed of sorting, high speed articulated robot and machine vision system comes to focus. In this inspection system, Allen-Bradley® MicroLogixTM Programmable Logic Controller (PLC) is designed as main controller that controls the MOTOMAN robot, Cognex [5] camera vision system, the servomotor, and ancillary units. The MOTOMAN robot is programmed using a different user frames such that the inspection criteria can be easily changed and implemented. Figure 1 represent a MOTOMAN HP3C robot mounted with Cognex [5] camera and pneumatic gripper. Figure 1: MOTOMAN HP3C robot mounted with Cognex [5] camera and pneumatic gripper Each of these user frames should assign the position variables which mean the X, Y, Z coordination of a fixed location of the system. A Cognex [5] camera has been trained to recognize a defect of a part and send the right digital pulse to the robot for separating the defective part from the main product line to reject conveyor. Here the communication protocol among PLC, Cognex [5] camera and robot are important because together they determine the defect and separate the unacceptable parts. This research will encompass study involving component sizing required for the retrofit of the existing error-prone inspection system. Both electrical and mechanical aspects will be covered in this paper, with emphasis on the robot programming, vision system training, PLC programming and other electrical aspect. Sizing components and overall system setup will be dictated by a proper understanding of the process and its requirements. The electrical scope of the project will cover frame by frame MOTOMAN industrial robot programming, Allen-Bradley® MicroLogix PLC ladder logic programming, training Cognex [5] camera vison system for pattern recognition, and organizing sensors and electrical system to work in synchronism for accurate inspection and separation of defective parts. This research will emphasize on the ease of adoptability and automated calibration of the inspection system as the defect location, defect type, part type vary. The mechanical scope of the project will cover End of the arm tool design, Robot cell design, robot base design, and modification of the secondary conveyor stands. This arrangement will represent an intelligent inspection and sorting setup used in the industrial application. This inspection system can modified rapidly to fit any unique characteristics that is required to be inspected for the product at hand. These inspection characteristics determine the ultimate quality of the manufactured products. The products which do not meet the required quality will be diverted to a reject lane for further inspection and recycling. The robot and PLC programming structure must be modular to provide scopes for future extensions and improvements. Problem Description: This inspection system is designed to inspect a small foreign object anywhere on a 4 inch by 3 inch part as per the trained Cognex [5] camera vision system. Before the inspection, the part is picked up by the robot from a specific location on the conveyor system and is placed on a backlighted table in order to achieve the higher accuracy of the inspection process. If the part passes the inspection process it continues on the conveyor. However, if the inspection fails as per the trained criteria the part is separated to the reject conveyor for further processing. The Cognex [5] camera is mounted on the robot arm and back lighted table is used to achieve the speed and accuracy of the inspection process. The camera is faced straight towards the top face of the product, a top view of inspecting part. Any change in the inspection criteria, change in the part type or change in part size can be adopted promptly in to inspection process for modular programming techniques. The camera is programmed based on the inspection criteria of the product. The inspection criteria can vary widely. Any mismatch with the pre-saved inspection pattern, the camera will detect it that is difficult to detect at high speed by naked eye. Based on the unique criteria of the product, parameter values of certain unique criteria will be extracted at the point of view of the inspection tool. Then the camera will compare with pre-saved formats and identify the product based on the comparison. Criteria for Accepted Product For the target setup we assume the criteria for an accepted product is “empty / blank holes”. Figure 2 shows a sample accepted product. This is analogous to a clean end product. Figure 2: Criteria for accepted product If the camera does not see any objects in the holes, the products will be detected as accepted. Only the red-marked region as shown in Figure 3 on the products will be inspected for the inspection. Empty / Blank Holes Figure 3: Region to inspect Criteria for Rejected Product Figure 4: Sample rejected product For our test setup we assume that the criteria for a rejected product are random foreign objects in holes of the containers. This object can be any random non-transparent visible object in the point of view of the camera. Any presence of any visible unknown object for example, pieces of nuts, screws will trigger a rejected result for the camera.