Students: Julius Bötger, Sarah Stam, Nathan Saeri & Jesse Starmans
TIJSSEN GOED VOOR DIEREN ANOUNCES AN AUTOMATIC PACKING ROBOT FOR BAGS OF ANIMAL FOOD.
Tijssen Goed voor Dieren has designed and made an automatic packing robot using a vacuum gripper. With this solution, it reduces the need for manual labour and increases the rate of production.
A SMR group has made their automatic packing robot for the animal bag producer Tijssen Goed voor Dieren. Using a combination of machine vision, a vacuum gripper and a robot arm, it is possible for the robot to locate, pickup and box up various sizes of bags of animal food in various configurations. This solution reduces the need for manual labour and thus the possibility of injury. Furthermore, this can speed up the production process.
Boxing up bags of animal food, ranging from a few hundred grams to several kilos, is a very intensive manual task. Because of this, it is not really possible for a person to be packing up bags continuously, for risk of injury. They also need to go home, so it is not possible to run the production non-stop, which is unused time in which the company could have been making more money. When fully optimised, a robot arm can also move faster and more precise than a human can. Due to this, the manufactured products can be boxed up faster.
To achieve this, a gripper had to be made that can form a good seal with a great variety of bag materials and sizes. This was solved by designing a gripper, made of two parts, that has an industrial vacuum pump attached. With this, it is possible to lift a small bag with only one of the grippers or a big bag with both grippers together.
This solves the problem of needing a person to pick up heavy bags and because a robot theoretically does not need to stop, the production can run as fast as possible for as long as needed.
To supply the robot with the necessary starting-information about the box in which the bags need to be placed, the height and weight of the bag and the amount of bags per box, a Graphical User Interface (GUI) has been made. With this GUI, it is also possible to start, pause and stop the robot, thus needing only one screen to control all the necessary aspects of the robot.
When a bag comes out of the production machine, a ultrasonic sensor detects it and tells the camera above to make a photo. A Object detection model is applied to this photo to recognise the bag and ascertain information about the exact position of the bag and the rotation. This is send to a backend server, which converts this data to data for the robot. The robot then uses this data to ultimately pick up the bag, put it in the box in the correct orientation and configuration and then waits for the next bag.