+++
We are at the Hannover Messe Hall 15, Booth H07.
Learn more
+++
Use case

Assembly

DeepInspect is a research prototype for a highly flexible and easy-to-learn camera- and AI-based assistance system for manual assembly stations.
Project partners involved

Assembly assistance for the perfect assembly guide

The DeepInspect system approach allows customers with many product variations and challenging manual assembly tasks to maintain or bring both quality and productivity to high levels. This is done by assisting the assemblers with the AI-supported software regarding the correct assembly, respectively by providing step-by-step assembly instructions.
Project partners involved

Problem:

  • Manual assembly is very stressful for people. Especially in case of high product variants, as well as in case of strong monotony, technical solutions can assist human beings during assembly.
  • At the same time, quality inspection is essential to meet quality standards.
  • DeepInspect aims to assist the assembly of the product with the help of modern object recognition methods. Both in terms of guiding the employee interactively, as well as in terms of checking the product for correct assembly.

Solution approach:

The Four DeepInspect Steps

  • Step 1: Capture and annotate images.
  • Step 2: Define correct position and part type.
  • Step 3: Define workflow and teach-in.
  • Step 4: Apply workflow correctly using assistance system.
Components

Added value through FabOS

While the system is designed a very simple and easy-to-use web-application to the user (web frontend), the overall system consists of several modular and distributed backend services, designed in a way, that they might be reused by different applications, additionally to the DeepInspect systems as such.

The outcome w.r.t. to FabOS research and development are:

  • modular system design components for image capturing, image streaming, object detection etc. provided re-usable and FabOS-compliant deployable units.
  • Design and implementation of an open (FabOS compatible) interface / API for interaction with further production systems to allow an orchestration of value chains (orchestrated production lines).

Component identification
Component identification solution that uses a neural network and vector embedding approach to quickly and accurately identify components in real time.
Learn more
Asset Administration Shell
Asset Administration Shells (AAS) are the basis of the Digital Twin with uniform interfaces for all types of production equipment.
Learn more
No items found.

Your contact for questions

We have received your request and will be in touch soon!
Oops! Something went wrong while submitting the form.

Discover more use cases

Stock

Stock

By combining the technologies of KENBUN, TRUMPF and FabOS, an intelligent warehouse with voice assistance can be realized that is robust and suitable for everyday use even under noisy conditions.

Measuring

Measuring

AI based solutions do support the processes in a versatile production beginning at the material provision with smart carrier boxes, over the set-up of intelligent workpiece carriers, the active temperature control of workpieces up to the temperature supervised measurement process.

Milling

Milling

Intelligent process monitoring makes it possible to ensure high precision and surface finish requirements on milled parts.

Laser cutting

Laser cutting

The removal of laser-cut components is usually done manually and is both time-consuming and tedious. Therefore, an automated removal solution is urgently needed, using suction plates and pins and optimized by AI models.

Bin Picking

Bin Picking

Gripping parts from crates is a common problem in many factories and presents significant challenges.