Industrial Automation branch INA 
Flexible robot grippers for recognizing, gripping and inspecting natural objects in an Industry 4.0 environment (I4KMU-FlexRoG)
Summary
In industrial production, the automatic assessment and sorting of components and products is a standard procedure. Depending on the complexity of the quality criteria and the nature of the products, the technologies vary in complexity and cost. The cooperation between Fraunhofer IOSB-INA and the special machine manufacturer Anlagenbau Habeck GmbH & Co. KG shows how quality control can be efficiently automated.
Challenge
It takes a trained eye to assess the quality of food.
In addition to shape, consistency, color and texture, many factors play a role.
The central challenge of the project was to develop an automated, robot-assisted quality control system that is also financially attractive for small and medium-sized enterprises (SMEs). Existing technologies designed for the mechanical manufacture of products or product components were not suitable for the target group and the specific application. The aim was therefore to create a low-threshold solution that draws its flexibility from the combination of different domains such as intelligent sensor systems, machine learning and innovative plant engineering. The high cost of implementing and programming robots is a major obstacle for SMEs. Robots require specific end effectors and the ability to react flexibly to changing production requirements, which further increases costs.
Objectives & approach
Exemplary application scenario: A vacuum gripper is to grip different organic objects and check them for quality characteristics in a single work step.
quality characteristics in a single work step.
The project "Flexible robot gripper for recognizing, gripping and inspecting natural objects in the Industry 4.0 environment (I4KMU-FlexRoG)" aims to significantly improve the programming effort, the mechanical limitations of the gripper and the functional flexibility through a key combination of robotics, deep learning and machine learning. Habeck is developing a specific robotic gripping system that can handle various gripping objects and check their quality during handling. Automated and process-stable object recognition is intended to reduce programming effort. The main area of application is the gripping and inspection of potatoes in the food industry, although the technology should also be transferable to other types of fruit and vegetables. This flexible, easily convertible technology is intended to quickly amortize one-off development costs for different application scenarios in order to offer an attractive solution for SMEs. The combination of Fraunhofer's research expertise and Habeck's development know-how creates a cross-industry solution for quality control in the food industry.
Results & values
Several key performance indicators (KPIs) were defined in the project:
- The flexible gripper should be combinable with articulated arm robots, regardless of manufacturer, in order to function as a retrofit solution for existing robot systems.
- Robot programming is to be simplified by position and object recognition, which independently determines the position of the workpiece.
- The gripper should be able to analyze workpieces as part of an inline quality control process so that robots can be used directly for sorting tasks.
An initial prototype of the gripper, which is based on the vacuum principle, has been developed. It can handle different workpieces, but has not yet been combined across manufacturers and does not have the necessary sensor technology. The gripper is currently being further developed so that it can be used with robots in Industry 4.0 environments. Sensors based on optical technologies are being integrated in order to realize manufacturer independence, position detection and inline quality control. A testbed at the SmartFactoryOWL in Lemgo serves as a development environment.
Customer benefits
The developed solution offers SMEs an economically attractive opportunity to benefit from the advantages of robotics. The flexible gripper enables rapid amortization thanks to its reusability in various scenarios and applications. In the food industry, where robots have so far mainly been used for transport, palletizing and packaging, the new technology can also be used for quality inspection. This increases efficiency and reduces costs. Thanks to the reduced programming effort and the flexibility of the gripper, companies can react more quickly to changing requirements. The solution promotes the implementation of advanced technologies such as deep learning and machine learning in SMEs, enabling them to benefit from the efficiency gains of Industry 4.0. The practical transferability to other products and sectors with low installation costs creates additional potential for industrial applications. Overall, the project offers a promising perspective for the introduction of cost-efficient and flexible quality control solutions in the food industry and beyond.