Automatic Bottle Unloader | Elkemelk

Student : |Daan Van der Klugt|Bastiaan Snoeij|Elena Gabriela Voicila|Ihsan Fadilah|(Ainur Rafika Firdausy)

ElkeMelk, a dedicated dairy company committed to producing high-quality milk. To ensure this premium milk reaches a wide audience, ElkeMelk continuously strives to meet the demands of consumers by maintaining consistent and sufficient production for distribution.

Automation plays a key role in the production of nutritious dairy products at ElkeMelk. Currently, a turntable is used to place bottles before processing. However, the task of unloading bottles onto the conveyor remains a manual process. These activities not only take up a lot of time but also require workers to be in the factory for a short period of time at a time. In addition, the company also faced the challenge of developing a technology to sanitize the bottles before filling them with milk as the labels on the bottles would melt after passing through the sterilization process. These ongoing issues show there’s still work to be done to get everything running smoothly.

To address these challenges, the team successfully designed a robotic arm capable of grabbing bottles, flipping them upside down for the cleaning station, and placing them onto the conveyor before labeling. Developing this solution required extensive brainstorming to ensure the robot could detect bottle cups accurately, extend its reach to access further rows, and operate at optimal speed.

 

  • End Effector

To improve the robot’s reach, a custom 25 cm end effector/gripper adapter was created. This end effector houses a camera and features four suction units at the base, with a gripper at the front. Foam padding was added to the gripper to ensure a firm grasp on the bottles.

 

  • Camera Vision

Accurate bottle detection required high-quality imaging. The team used a gaming camera capable of capturing high-resolution images to enhance detection accuracy for transparent bottles. The YOLO v8 model was implemented to detect bottle caps, allowing the robot to identify all bottles on-screen, determine their coordinates, and locate the center point for precise handling.

 

  • Logic

To maximize the robot’s efficiency, the joint rotation was optimized, and the carton area was divided into two parts for handling. The first part (Part I) involves grabbing the first four rows from the side, and once completed, the robot transitions to Part II, which involves grabbing bottles from the front. This strategic division allows the carton layer to be placed close to the robot without interfering with its operation when handling the initial rows.

Current Results

The current results demonstrate the robotic arm’s remarkable reliability and efficiency in meeting production targets. With a consistent processing rate of three bottles per minute and flawless performance in grasping and moving each bottle, the robot achieves the precision required for a streamlined workflow. Its ability to seamlessly bypass any missing bottles and continue to the next ensures an uninterrupted process, highlighting the robustness of its design and programming. Moreover, the system’s adaptability allows it to maintain its high level of accuracy even in low-light conditions, further enhancing its versatility and reliability in various operating environments.