NeuraConnect: Outsourcing compute-intensive operations to a 5G edge cloud
Initial situation
Neura Robotics GmbH's high-tech robotic solutions enable humans to express their creativity and focus more on social interactions instead of doing unhealthy and monotonous work that is still very common in society. Neura Robotics GmbH is closing the gap in human-robot interaction by enhancing the cognitive capabilities of robots through revolutionary breakthroughs in environmental perception, mechanical design and artificial intelligence, enabling anyone to control and use a robot without any special prior knowledge.
Neura's robotics solutions can see, hear and feel their surroundings. Complex AI algorithms are used for this, which place high demands on the robot hardware, particularly in the areas of image processing, object recognition and the associated artificial intelligence.
Solution idea
NeuraConnect aims to connect robots and infrastructure so that external resources can be used to evaluate the complex AI algorithms. The solution idea is to outsource computationally intensive operations to a 5G edge cloud.
The prerequisite for this is a high-performance network connection, as is guaranteed in modern 5G networks. This means that large amounts of data in particular, such as image and audio data, can be evaluated centrally in real time.

mobile robots, source: 5G4KMU Transfer Center
Benefit
The growing robotics market means that several robots will be operating simultaneously in the same environment in the future. 5G has two advantages here:
Implementation of the AI application
Implementation and evaluation of a typical setup from Neura Robotics as a proof of concept (PoC) to validate the 5G transmission performance of image data. Fraunhofer IPA has a 5G test environment at its Stuttgart site. Various methods for data transmission and reference measurements for WLAN are carried out for the evaluation.
An analysis of relevant performance values such as latency, jitter and upload and download speeds is carried out and compared.