Object recognition technology is helping to employ robots in traditionally difficult cleanroom bin-picking operations.
Visual intelligence, or object recognition technology, is the ability of robots to identify different items, based on their three-dimensional geometry. Automating processes this way represents an effective way of increasing productivity and reducing costs on a vast range of general material handling applications.
Capable of locating parts three dimensionally, it adds flexibility and reliability to operations traditionally completed by humans or which otherwise require sophisticated and expensive dedicated machinery.
Typical applications include: de-palletising materials (including mixed boxes, sacks, bags and food packaging; bin picking (including loose random parts, irregular items, and irregularly-shaped sacks and packaging; sorting, placing and loading picked items into machines. A range of grippers is available to handles a vast range of materials, based on mechanical, magnetic or vacuum principles. Even setups involving dirty, dusty or rusty products and/or difficult light conditions can be realised.
There are several different ways to implement 3D vision applications:
* Stereo vision
* Stereo vision using structured light
* Laser profiling
* Time of flight
Moreover, the diversity of existing sensor technologies spans a wide spectrum of resolution and cost. Once the data is obtained from any 3D sensor, it needs to be analysed to provide location and orientation information that will make sense to an industrial robot.
FANUC‘s high-speed 3D Area Sensor is an example of a system which uses structured light projection to create 3D maps of its surrounds. Using these maps, the system looks for parts. The part manager then does an evaluation and decides which part to pick. Taking reaching distance and collision avoidance into account, it then chooses the fastest picking option. If the part manager decides a pick has been unsuccessful or a part queue does not contain a part to pick, another image is taken and the process starts again using the new results. Multi-tasking background processing – part detection takes place while the robot is moving and does not interrupt the workflow- means that shorter cycle times can be achieved.
In a typical cleanroom application, a robot equipped with the iRVision 3DA/400 Area Sensor can locate and pick randomly orientated bottle caps from a bin. “FANUC’s latest high-speed 3D Area Sensor is ideal for bin picking applications, which has traditionally been a very challenging robotic process,” said Bernhard Walker, material handling engineer, FANUC America.
The 3DA/400 Area Sensor provides 3D location of the bottle caps in the bin. The robot picks the bottle caps from the bin and places them in a second bin at high speeds. iRVision’s Interference Avoidance feature prevents the robot and tooling from coming in contact with the bin walls. DCS Speed and Position Check Software restricts the travel of the robot to the exact area in which it works.
The family of LR Mate 200iD robots is versatile in manufacturing operations that require access into small spaces. The 200iD/4SC has an ISO Class 4 (Class 10) cleanroom certification for electronics, pharmaceutical and food handling applications. It features a white FDA compliant coating, stainless steel wrist, and NSF-H1 grade grease on all joints to provide reliable performance in demanding production environments, including rigorous sanitation procedures. The high-speed LR Mate 200iD/4SC has a maximum speed of 4000 mm/s, and includes a bottom cable exit option, making it ideal for applications where a small footprint is required.
A very slim arm about the same size as a human arm, and a bottom cable exit option minimize interference with peripheral devices. The LR Mate 200iD robots offer a “best in class” work envelope for both upright and invert mount installations. The LR Mate 200iD robots are also available with ISO Class 4 clean-room and food-grade variants for primary (unwrapped) food handling and healthcare packaging applications.
The 3DA/1300 Area Sensor can be top-mounted on an auxiliary axis-powered rail, allowing the robot to directly control the movement of the sensor. Mounting the sensor this way gives the robot two bins from which to work. As soon as the robot recognises that one bin is empty it will automatically switch to the next one, saving downtime that would be required for an operator to change the bin manually.