The result is a procedure to secure arm motions obtained by a learning by demonstration techniques, by adding a supervised Machine Learning classifier to detect anomalous arm positions. The main result is that this solves the burden to obtain large datasets, and this can be re-run for each motion that is learned.
Addressing target audiences and expressing needs
- Use of research Infrastructure
- Collaboration
This result is a result whose scope is twofold: on the one hand it is ground for active research both in the robotics field and also in the AI field. Also applications of these results are of interest, namely in the context of the evaluation of Human-Robot Interaction.
On the other hand, while collaborative robots are still products for research, end-users of such technology are also seeked. The interest to go for a human-centered Industry 5.0 has to be done by means of successful pilots in real SME.
- Research and Technology Organisations
- Academia/ Universities
- Private Investors
R&D, Technology and Innovation aspects
This has been tested in the collaborative Robot TIAGo in the PAL Robotics premises, both in simulation and with real machines.
The attached link to the paper provides a descrption of the explicit procedure to produce the exposed trained machine learning models.
Having in mind the effective demonstration of the viability of cobots on the shop floor, PAL Robotics is currently manufacturing the three TIAGo research platforms that are going to be used as the collaborative robot (cobot) of the SeCoIIA project in the context of the cross-sectorial use-case of the project. PAL Robotics is working on a fleet management system to be able to deploy and control different robots at the same time on the shop floor. Moreover, a software package to handle anomaly detection has been developed and made available to the consortium to include their machine-learning based trained models to aid the safe and secure deployment of the cobots in the shop floor.
Autonomous Mobile Robots (AMRs) are an easier, more affordable and versatile solution than Automated Guided Vehicles (AGVs) in the shop floor and perform the repetitive tasks to be able to adapt to a changing environment.
PAL Robotics’ platforms run under the standard tools used by robotics companies nowadays.