| Overview | Projects | Publications | Archives | Facilities | Directory |

SPIE Web's (International Society of Optical Engineering) OE Reports - September 1998

Robots hired in poultry packing plants

The Intelligent Integrated Belt Manipulator (IIBM) robot tackles a common food industry task by removing items from a conveyor belt and transferring them into a packing carton for shipping. It is undergoing field testing at a ConAgra poultry plant in Gainesville, Georgia.

Many jobs in poultry processing consist of materials handling tasks, such as moving a product from a conveyor belt into a box or onto another conveyor belt. Gary McMurray, a senior researcher in Georgia Tech's Agricultural Technology Research Program (ATRP) and a project director for the robotics initiative, said that their goal was to develop a low-cost robot that could perform materials handling tasks with the same speed and dexterity as a human. Although poultry plants are using simple forms of fixed automation, these machines have very limited capabilities and are expensive.

Georgia Tech researchers have been developing a new breed of robot that will increase efficiency and competitiveness for the poultry industry. The Intelligent Integrated Belt Manipulator (IIBM) tackles a common food industry task by removing items from a conveyor belt and transferring them into a packing carton.

Conceived in 1992, IIBM has gone through several redesigns and refinements over the years. The first-generation robot was powered exclusively by pneumatics, attractive because of its low costs and ease of use. Noting that the prototype fluctuated up to an inch in position when picking up items, McMurray said that the speed was good but that the accuracy rate did not meet expectations -- the robot picked up the product, but sometimes misplaced it in the shipping carton.

Poultry pieces differ considerably in size and shape, varying the contours of the packages by up to two inches and causing a shift in weight and center of gravity. This made grasping demands another challenge for the IIBM, and its end effector had to be constructed with some flexibility. Suction cups were made from bellows material (which compresses up to three-quarters of an inch), and a spring mechanism was added to those suction cups providing another inch of compliance. These changes allow the grippers to conform to different contours of a product.

After four months of lab testing, the current IIBM prototype has been sent to the factory floor in a ConAgra plant (producers of Butterball and Country Pride poultry products) in Gainesville, GA to focus on speed and accuracy testing. Early field test results have been encouraging. In lab trials, the IIBM's average cycle time was clocked at 2.1 seconds -- comparable to a human worker. This timing has been sustained in the plant as well. During lab testing, the robot occasionally dropped a tray pack, but rarely missed picking up the product.

The new IIBM is a hybrid of pneumatics and electro-servo drives. Two pneumatic axes and two electro-servo axes allow motion in four directions -- up and down, parallel to the conveyor belt, perpendicular to the conveyor belt, and at a 90-degree rotational pivot.

Besides cutting costs, the IIBM is attractive because it is simple both to install and to maintain. Only a few physical dimensions must be programmed -- the size of the tray packs, the location of the packing cartons, and the height of the conveyor belt.

J. Craig Wyvill, director of ATRP, hopes to have the technology commercialized within the next two years. His next step is to enhance the system with a vision system, which would develop hand/eye coordination for the robot and allow it to operate by merely seeing a picture of the product, eliminating the need for task-specific software and programming.

McMurray stresses that the IIBM wasn't created to perform just one materials handling task, but rather to demonstrate the cost-effectiveness and flexibility of automation for the poultry industry. He estimates that final commercial costs will range between $30,000 and $40,000 -- about half the price of existing industrial robotic systems. Preliminary discussions are already being held with private companies.