SUMMARY REMARKS & TESTBEDS
CAFA Tech is developing autonomous wheeled robots RAM (Remote-Autonomous-Mobile robot) with max payload 150kg and speed up to 15km/h for Logistics, Construction and Agriculture companies. The RAM-20 robot is designed for logistics tasks and to drive autonomously indoor and on outdoor limited (in private premises) territories, as well as on sidewalks and on the streets.
By September 2020, during Fed4Fire Phase Stage 1. Small experiments have tested the location of the CAFA RAM-20 robot indoors, using markers detected by the robot’s camera and GPU-based computer vision software. The next step is to develop the RAM-20 robot self-driving function for outdoor driving. For self-driving RAM robot must know information about humans and other moving objects (cars, trucks, robots, rollers etc.).
It is necessary to experiment with GPUs of different capacities, to identify which tasks can be performed on the robot onboard and which edge server. GPULab has different GPUs and settings to support these experiments. When driving on the streets, it is important to identify other vehicles and transmit the location and direction of the CAFA RAM-20 robot to other vehicles and to receive information from other (e.g. from self-driving cars) about their trajectory and location in near real-time. In EU there are two main solutions for V2X (Vehicle to everything) communication at 5.9GHz: ITS-G5 and C-V2X. During the project CAFA Tech plan to use CityLab Smart Highway testbed which supports the use of both standards on the campus of the University of Antwerpen and on the street).
CAFA Tech intends to perform following Stage 2. Medium experiments: June 01-Oct 31, 2021:
- Computer Vision solution experiments to identify outdoor markers, obstacles, people and vehicles close to the robot in near real time (experiments in GPULab).
- V2X solution experiments to transfer information from the CAFA RAM-20 robot to other vehicles and vice-versa (experiments in CityLab).