Mobile Reutlinger TORY (MoRTy) is our newest robot family member.
He is already working on our project “AI-on-the-edge” and the deep learning-based “Follow-Me-Application” for service robots.
The most important advantages of the innovative application are listed below:
(i) First implementation of a robust follow-me function in dynamic, busy environments with groups of people for a service robot.
(ii) Highly precise navigation through the new deep learning method based on a sensor fusion and monocular non-rigid slam method on ai-on-the-edge technologies.
(iii) Ai-on-the-edge systems achieve a reduction in bandwidth capacities, online times, latencies, etc., since the necessary calculations are carried out directly on the service robot instead of via network connections “in the cloud”.
(iv) Reliable and unambiguous identification as well as tracking of people and re-identification are based on self-learning AI processes.
(v) Development and implementation of the collision-free movement trajectory and the tracking system based on a new visual slam process and a combined machine-vision process.
(vi) In close cooperation between the project partners, the functionality of the innovative application can then be validated in practical tests and finally optimized.
For this project, we set up the research platform TORY. It is differentially driven, has a 40Ah LiFe battery and a safety drive control with the safety functions SLS, STO and ST1 to meet relevant safety requirements.
In terms of sensors, the platform is therefore equipped with a SICK S300 2D safety laser scanner, a SICK TIM 551, two solid-state lidars for obstacles close to the ground. There is also an Orbbec Astra monocular RGBD camera for 3D obstacle avoidance in the SLAM process.
According to the current results from work package 3, an NVIDIA Jetson AGX Xavier and, as an option, an Intel Movidius Myriad X are prepared for the application of processes with artificial intelligence on AI-on-the-edge devices.
Read more about our robots
