Transportation is essential part of production: Flexible, programmable transport, and dynamic adaptation are needed due to changes in requirements, environment, or operation policy.
Decoupling control from HW: Remove all control from the robot and execute at central place (edge cloud). This approach allows for simpler and more cost-effective robot architectures, with the hardware and software components evolving independently. It also enhances resilience against software and hardware failures, while facilitating software maintenance tasks.
The introduction of 5G capabilities opens up new possibilities for various use cases, such as handling computationally intensive tasks and executing tightly synchronized operations.
Extend the currently available HW and SW platform (two AGVs and their SW stack) to demonstrate 5G-enabled new capabilities such as collaboration and real-time execution of image processing in the edge. Implement use-cases for this purpose such as:
Camera-based real-time and highly-accurate self-positioning and localization,
Collaboration of AGVs using social layer information from camera input.
Reduce video coding/transport/processing delay components.
Compensate video delays (~200 ms) with AI prediction using zero-delay additional sensors and control info.
Measure application KPIs such as positioning and localization accuracy.
Demonstrate the main features of the proposed mechanisms.
Dániel AGÓCS, Marcell BALOGH, Gábor FEHÉR, Dániel RÓZSA, Attila VIDÁCS
Fehér, G., Rózsa D., Balogh M., Reider N., Extending Edge-Based Mobile Robot Navigation with Social Awareness. (2023) International Symposium on Industrial Electronics (ISIE), Helsinki, Finland https://2023.ieee-isie.org/program
Fehér, G., Agócs, D., Rózsa D., Balogh M., Reider N., Compensating Delays for Precise and Real-time ROS Cloud Robotics Localization. (2023) International Symposium on Industrial Electronics (ISIE), Helsinki, Finland https://2023.ieee-isie.org/program
Exhibition Demo at AI&Aut Expo - Exhibition for Local, Regional and European Technology and Innovation, February 2023, Budapest, Hungary
We introduce a detailed realization of industrial mobile robot navigation abilities among human workers. We extend the navigation layers with state of the art person detection and adapt them to influence trajectory planning. We present our cloud-based architecture, how onboard components are moved to the cloud, and how person detection is integrated into navigation. The system also incorporates social awareness, allowing the robot to take into account the presence and actions of people in its environment and adjust its navigation in line with the requirements of industry 5.0. The realization is evaluated in a series of experiments where results show that it is able to effectively navigate in industrial environment with people present and avoid collision with people or other objects.