TIAGo robot by PAL Robotics is suitable for perception, navigation, manipulation & Human-Robot Interaction. Different versions of TIAGo are available, including TIAGo Lite (no arms), TIAGo (single arm), and TIAGo++ (dual arm). The robot is a popular choice for research in universities and innovation centers, especially in Europe but also around the world, as well as being used in a number of collaborative EU projects.
In recent years, the team at PAL Robotics has been working hard to further develop TIAGo’s capabilities in order to incorporate some of the most sophisticated mobile manipulation technology available today, and to meet the needs of researchers and industry.
The version of TIAGo used in PILLAR-Robots is known as TIAGo SEA. This robot features an omnidirectional base that moves in all directions and is especially useful for movement in constrained spaces. In addition TIAGo SEA includes state-of-the-art arm design which incorporates the following features:
- Torque control
- SEA actuators
- New grippers for object manipulation
Series Elastic Actuator (SEA) arms are a groundbreaking advancement in robotic technology. Through their design, they provide dynamic grasping, better vision-based manipulation, dexterous assembly, haptics, deep learning, mobile manipulation and Human-Robot Interaction capabilities. The SEA arms offer safer and lighter robots, interchangeable end-effectors, and a wide array of upgradeable sensors.
PILLAR-Robots use cases
PILLAR-Robots will implement and validate demonstrators within three different application fields characterised by different types and levels of variability: Agri-food, Edutainment, and unstructured Industrial/retail. Each of these demonstrators includes the TIAGo robot by PAL Robotics.
Here, the TIAGo software and cognitive architecture (and potentially hardware, for example, adapting the gripper if needed), will be adjusted. A user interface will also be developed to allow non-specialised operators to change the robots’ purposes.
To start with, a simulated environment allowing the emulation of typical picking/sorting/packaging activities will be realised. For this, PAL Robotics will furnish the TIAGo model definitions suitable for a PyBullet-based high-speed simulation as needed by open-ended learning. The testbed will involve challenging features such as different fruit aggregations (individual, grouped) and kinematics (fruits/vegetables on a still plane or on a convey belt).
Secondly, a lab-engineered environment using plastic or natural fruits/vegetables will be created, and AI2L will perform the experiments using the CNR’s TIAGo robots. The robot will refine the skills acquired during the simulated learning phase to work in the physical environment.
Lastly, a reduced-complexity environment located in an industrial post-harvesting factory will be developed, and here TIAGo will learn to manipulate natural fruits and vegetables within a real industrial context, by refining the skills acquired in the lab-engineered environment.
The project will introduce this demonstrator in non-formal education environments, such as extracurricular activities in schools, museums, adult training and care contexts, with the aim of assessing the capabilities of this technology to adapt to two hardware platforms.
The first validation stage will be carried out with the Robobo robot, a suitable robot for younger audience in part due to its pet-like appearance. The second validation stage will involve a TIAGo robot, with advanced manipulation capabilities.
To carry out this validation, the project will first rely on a simulated environment, then on a lab environment, before testing in a realistic experimental facility.
Unstructured Industrial/retail Scenario
The target here will be workshops and warehouses that have not been built for automation. TIAGo robot will therefore need to deal with large variability, both in the environment and in the tasks. This variability requires a strong involvement by programming experts that will be replaced by the proposed purpose-based intrinsically motivated open-ended learning approach.
The project will carry out this validation first in a simulated environment, then in a lab environment, and then following this, by testing in a realistic experimental facility.
About PAL Robotics
PAL Robotics is based in Barcelona and has the mission to enhance people’s quality of life through service robotics and automation technologies. PAL Robotics design and manufacture highly integrated and reliable solutions for research institutions and service industries worldwide.
PAL Robotics started in 2004, when a small group of engineers built the first fully autonomous humanoid biped robot in Europe. With many years of experience in R&D, today they are known for their humanoid and mobile robots used for research, logistics, retail and social applications for companies, institutions and labs.