Fixed automation systems and conventional blind robots have a role in automated assembly, but it’s limited. Fixed systems are preferred in applications where very high volumes of one part or a few similar parts are produced. Conventional blind robots are often used where parts and tools need to be located within close tolerances.

Vision-guided robots, in contrast, enable manufacturers to automate a wide range of applications—including processing small batches of parts or mixed batches of multiple parts; locating parts with a very high level of accuracy; and picking and assembling parts regardless of their location as long as they are within the vision system’s field of view.

State-of-the-art vision systems maximize robotic capabilities by interfacing directly to robot controllers, PLCs, HMIs or PCs. They also incorporate Ethernet ports that enable them to be directly connected to any switch or hub on a factory network and, in turn, communicate with all other devices on the network.

Aware of these benefits, manufacturers from various industries are increasingly using vision-guided robotics to automate their assembly processes. The following case histories showcase four of these manufacturers. For all of them, the key to suc-cess is matching the right camera and vision software to the specific application.

Bin Picking and Assembly

One of the most difficult and time-consuming processes performed by oil tool manufacturers is installing thread protectors onto exposed threaded pipe. Over the years, many manufacturers have tried to automate the process, but failed because the systems were not able to retrieve many different sizes and styles of protectors and pipes.

That is no longer the case, thanks to JMP Engineering. This Canadian integrator recently developed a flexible automation process that enables an oil tool manufacturer to handle multiple-size protectors and pipes.

The process uses two robots to install protectors on a wide range of pipes, but can be easily configured to handle future variants without programming. Mounted on the arm of each robot is a Basler Ace camera, which has a frame grabber card and is connected to a Beckhoff industrial PC. JMP programmers also wrote a graphic user interface (in Visual Basic) that performs vision operations by calling up tools from the Cognex VisionPro software library.

Thread protectors come in 11 different sizes (ranging from 4 to 8 inches in diameter) and are packed in bins in layers divided by cardboard sheets. Before a protector can be removed from a bin, the camera on robot one must use the PatMax tool to identify and calculate the protector’s location.

PatMax works by matching a set of boundary curves (not tied to a pixel grid) with similar shapes found in the camera image. When a match occurs, the PLC in the workcell directs robot one to retrieve the protector.

Shortly thereafter, robot one hands off the protector to robot two, which brings it to a fixture holding a section of pipe. The camera on robot two uses the Circle tool to determine the location of the pipe and verify its diameter to make sure it matches the thread protector. The robot then assembles the protector and pipe.

“There’s a good chance this application will lead to a new generation of vision-enabled robots that will improve productivity and quality in the oil tool industry,” says Scott Pytel, project manager at JMP Engineering. “These robots have demonstrated the ability to successfully pick and assemble thread protectors without fixturing or accurate locating in conditions that are common in oil tool manufacturing.”

Part Identification and Modification

Foundries tend to manufacture a variety of parts, and these parts often vary greatly in size and geometry. As a result, it has been difficult for foundries to automate their production processes.

Bucking that tradition is Odyssey Machine Co. Ltd., a foundry based in Bowling Green, OH. In 2011, Odyssey worked with RAF Automation, a systems integrator in Solon, OH, to develop a vision-guided-robot workcell that automates the removal of sprue, filter and flash from eight different aluminum castings.

The workcell features a conveyor, sawing and milling equipment, a Fanuc S900W robot and a Cognex In-Sight Micro 1100 vision system mounted above the conveyor. The vision system provides an image resolution of 640 by 480 pixels in an enclosure of just 30 by 30 by 60 millimeters. It features a non-linear calibration tool that enables mounting at angles of up to 45 degrees from the object to be inspected. The system also supports Power over Ethernet to provide power and network communications on a single standard cable.

A worker places each casting on the conveyor. As the casting enters the workcell, the vision system’s sensor uses the PatMax tool to verify casting type and location. PatMax does this by analyzing each casting’s geometric information (angles, size and overall appearance) and spatial relationships.

Casting data is communicated to the robot controller, which positions the robot’s wrist at the correct angle so it picks up the casting and places it on a linear saw to cut off the sprue and filter. The robot then moves the casting to a helical-end mill for flash removal, and places the finished casting back on the conveyor to move it out of the workcell.

“The workcell has successfully automated what used to be a difficult manual task with the danger of repetitive stress injury,” says Ron Leroux, president of Odyssey Machine. “It eliminates two of the three people previously required to operate the workcell. The automation portion of the cell is about $75,000, and the total cost of employing a person in a foundry is also about $75,000. So by freeing up two people, the robot quickly pays for itself.”

Vision-to-Robot Calibration

Tightening lug nuts that hold the wheel to the hub might seem to be one of the simpler aspects of building an automobile. But it has proven to be one of the most difficult to automate. The problem is that, typically, the vehicle is only roughly positioned by a conveyor and the wheels themselves are free to rotate, tilt and turn. Therefore, an ordinary blind robot is never able to find the nuts.

Automation specialist Radix Controls of Oldcastle, Ontario, has successfully automated this application by using a dual smart camera vision system to determine the position of the wheel—including its fore, aft, side-to-side positions and three rotational axes. With this information, the robot can easily move the nutrunner into the exact position and tighten the nuts. Automating this application made it possible for two people to move from difficult and stressful jobs to more proactive positions in the company.

The ability to quickly calibrate the robot is critical in this application because of the potential for the vision system to be bumped by moving equipment. Radix has developed a process that calibrates the robot in less than one minute.

After placing the lug nuts, the operator actuates the calibration command on the In-Sight 5403 smart cameras to determine the centerline position of the wheel.
These coordinates are sent to the robot controller.

The operator then moves the robot into position in front of the wheel, and the vision system determines the offsets between its own and the robot’s coordinate systems. Once the offsets are entered into the robot controller, the vision system and robot are synchronized.

Next, the robot swivels its wrist to match the wheel’s tilt and rotates the nutrunner to match the wheel’s angle of rotation. The robot positions the nutrunner squarely onto the lug nuts before cycling the nutrunner and tightening the nuts to the proper torque in a few seconds.

“The key to successfully automating this application is the coupling of machine vision and robotics, which accurately and repeatedly guide the robot to the proper position,” says Shelley Fellows, vice president of operations for Radix Controls. “Automation provides a substantial cost savings to the automobile manufacturer and also improves quality by ensuring that the lugs are repeatably tightened to the proper torque.”

Communication and Integration

Pharmaceutical manufacturers tend to use dedicated machinery based on fixed automation to fill vials, syringes, IV bags and other medical containers. Consequently, the machinery is capable of filling only one type of container.

This is acceptable for pharmaceutical manufacturers that produce long runs using a single container—but unacceptable for manufacturers that also need to package the same drug in multiple types of containers, produce short runs for clinical trials or produce smaller-scale personalized medications rather than big-selling drugs. For these applications, a manufacturer must have the flexibility to provide unique drug delivery systems tailored to its products and patients.
Machine designer and manufacturer Automated Systems of Tacoma Inc. (AST) recently helped one life science research company achieve this flexibility.
Specifically, the company wanted AST to develop a machine with a single platform flexible enough to fill and seal all their containers (prefilled syringes, vials, cartridges and IV bags) with low changeover times.

Such a machine would save the company money and clean room space by eliminating the need for upstream equipment that washes, sterilizes and removes endotoxin from the containers and stoppers prior to filling. The containers hold a wide variety of orphan drugs, vaccines, and general and specialty medicines.

Known as AseptiCell, the machine features a Cognex In-Sight Micro 1100 vision system mounted on the arms of two Staubli TX-60 HE six-axis robots. AST selected the Staubli TX-60 because of its ability to withstand aggressive cleaning and bio-decontamination required for the application.

Tubs of containers and stoppers are located within the operating envelopes of the two robots. Inside the tubs are fixtures that hold each container and stopper in an upright position.

During production, the vision system precisely locates each container and stopper and sends this position information to the robot controller. The robots, in turn, position their end-of-arm tooling, fill the drug containers and seal them with the stoppers.

The machine can handle a wide range of container types and sizes. Changeover only takes about 30 minutes. The workcell operator simply uploads a new robot program, replaces the container carriers and instructs the robots to change out their end-of-arm tooling.

Although used with a Staubli robot in this application, the Micro 1100 comes equipped with preconfigured drivers, ready-to-use templates and sample code that enable it to communicate with robots made by ABB, Adept, DENSO, EPSON, FANUC, Kawasaki, Kuka, Motoman, Mitsubishi and other manufacturers.

“Integration between the vision system and robot was key,” says Brian LaFave, system specialist for automation provider Olympus Controls, which helped install the vision system. “Mounting the system on the robot arm also made it essential for the system to be small, light and have very simple cabling.”