Industrial Sensors

The automation of production requires reliable systems for the real-time monitoring and control of processes. We envision Artificial Intelligence (AI) solutions that enable humans and machines to work together safely by detecting anomalies and preventing potential hazards, thus ensuring precision and reliability. To achieve this, we focus on the following two areas in the context of industrial sensor systems.

Contact us now

Dependable Person Detection

Object and Person detection through tablet

Robots have been an indispensable part of our industry for decades. They help in industrial assembly, as driverless transport systems in logistics, and in manufacturing in general. But high safety requirements often keep robots and humans physically separate, so today’s robot systems are typically fenced off in production areas. Safety fences, doors, and light barriers ensure the safety of human workers, but they also limit the efficiency of the systems.

But this will change in the future: The goal is to have collaborative robots with reliable AI-based person detection that can work directly with humans in the same workspace, for example in manufacturing. To ensure safe integration of AI for person detection in all kinds of situations, we integrate additional monitors such as body part detectors, as well as adaptive prototype-based detectors that can reliably handle false positives. In addition, we perform a systematic safety analysis for such a person detector in production with reasonable guarantees.

This allows for maximum efficiency and flexibility without additional delays due to false alarms, increased flexibility, and safety for human workers in production processes.

Our core competencies

Within the area of Dependable Person Detection, we concentrate on the following themes:

Person and Object Detection

Robust AI: Uncertainty Estimation

Standards for AI, Safety and Automation

Flexible Quality Inspection

AI-based automation has the potential to help companies perform visual quality inspection more efficiently and accurately. This is especially important in times of skills shortages and low-volume production.

Although Deep Learning (DL) solutions for image recognition have been providing breakthroughs for more than 10 years, these approaches have not yet been widely adopted in industrial environments. The inability to collect sufficient data on defective samples is a major obstacle to robust AI solutions. In addition, the low reliability and lack of transferability of DL systems to new products and changing environments are major challenges for practical use.

Fraunhofer IKS offers innovative solutions for data-efficient, reliable, and flexible visual quality inspection systems. Contact us to learn more.

Our core competencies

Within the area of Flexible Quality Inspection, we concentrate on the following themes:

​FAST -Feedback-guided Automation of Sub-tasks

​Modular Concept Learning

Blog articles on industrial sensors

Read more about Industrial Sensors on our Safe Intelligence Blog:


Visual inspection / 3.7.2023

FAST: How less data leads to early and reliable automation through AI

Visual information is often used to make important decisions. Automating these decisions requires complex AI systems, which are generally unable to make reliable predictions. FAST attempts to provide an answer to these problems.


Concept-based Models / 17.4.2023

How visual concepts help to understand an AI’s decision

Modern AI models have demonstrated remarkable capabilities on various computer vision tasks such as image classification. Their decision-making process, though, remains mostly opaque. This can be particularly problematic for safety-critical applications. Fraunhofer IKS explores ways to increase the transparency of this black box.


Industrial Automation / 15.2.2024

Laying the Foundation for a New Manufacturing Paradigm

As production technologies continue to evolve, it can feel like companies are permanently playing catch-up. Read on to understand the current key changes in industrial automation and why it is worthwhile to embrace them.


Out-of-distribution detection / 6.4.2022

Is it all a cluster game? A sneak peak into OOD detection problems

Can deep neural networks in image processing reliably recognize new, unknown test patterns? First considerations for different methods — especially for safety-critical applications.


Autonomous driving / 17.11.2021

Reliable detection of pedestrians in road traffic

The system in autonomous vehicles must be able to reliably detect pedestrians. Deep learning approaches are the main method used for this task. However, in comparison with classic software, the results of these approaches must also be checked and verified, which requires various complicated technical measures. This article provides an overview of these measures.


Automated Guided Vehicles / 5.8.2021

How robots in warehouses operate safely and efficiently

A simulation designed by Fraunhofer IKS paves the way for robots and humans to interact safely — without reducing efficiency.