Development and implementation of new methods for localization, map construction and motion planning using RGB-D sensors in robotic flexible manufacturing systems
Project title
Development and implementation of new methods for localization, map construction and motion planning using RGB-D sensors in robotic flexible manufacturing systems
Name of Beneficiary/Beneficiaries
Poznan University of Technology
Name of programme
National programmes
Competition
Leader VIII
Project value
PLN 1 198 705.00
Funding value
PLN 1 198 705.00
Project delivery period
from 01.01.2018 to 30.06.2021
Meet our team
See the result of our work
Autonomous driving between workstations
Demonstration of robot operation
What problem does our project solve?
The aim of the project was to develop new perception methods using 3D sensors to enable flexible manipulation robot operation in the manufacturing industry. The project developed a manipulative mobile robot capable of autonomous movement between manufacturing stations. This allows the robot to be used to work on several jobs simultaneously. Because cooperative arms are used, the robot shares the workspace with a human and there is no need for additional workstation preparation. In addition, the robot is equipped with 3D sensors (RGB-D cameras and laser scanners) to build a 3D model of the environment and manipulate objects, without the need for specialised robot programming. Thanks to the new perception system, the robot is able to locate and move between workstations, position relative to the workstation, identify and manipulate objects on the stage, avoiding collisions with machines. An interface for issuing voice commands to the robot has also been developed, which allows the robot to be reprogrammed in a flexible production system without the need for knowledge of robot kinematics and control.
Who is going to benefit from the project results?
Co-operative mobile-manipulation robots are the perfect answer to the labour needs of businesses. Particularly high interest can be observed on the part of small and medium-sized enterprises. Most large production facilities are already robotised using classic industrial robots. Our aim is to create a product (robot software) designed to run on robots working in a flexible production system, i.e. aimed at small and medium-sized enterprises currently responsible for more than 50% of GDP.
The price of new robots is becoming increasingly competitive, given the rising cost of human labour and the reluctance of humans to perform a certain class of tasks. Sensors are also emerging that, with the development of appropriate algorithms (declared in the design), can be used to create flexible manufacturing systems.
We refer to three demonstration scenarios in the project:
- low-volume industrial production in the automotive and construction sectors (robot-assisted operation of CNC machines for the production of aluminium and steel parts),
- the confectionery industry (production of buttons),
- laboratory tests performed by the robot (assistant in the chemistry or biotechnology laboratory).
However, the robot that has been developed is designed to perform most of the tasks that humans perform in the production process. With the right changes to the software, the robot can take care of tasks such as:
- sorting fruit (hard apples and pears),
- furniture production (for glue application),
- quality control (once the robot is equipped with the appropriate sensors), etc.