Trajectory Optimisation and Robotic Sequencing for Industry 4.0
When a set of parts must be transported from their current positions to their destinations during a manufacturing process, the order in which the available resources are used to carry out this task can be optimised to reduce the distance travelled or the time taken. To this end, algorithms have been designed to determine the optimal order in which a task should be performed, taking mechanical and/or energy-related constraints into account.
TECHNOLOGY DESCRIPTION
Many industrial processes require parts to be grouped or sorted during the manufacture of certain products. Typical examples include grouping the different parts that make up a product sold in disassembled form and intended to be assembled by the customer, or, in the fashion industry, grouping components of different shapes or materials that make up a product (textile garments or footwear).
In mass production, the order in which each item is transported results in productivity gains or losses in industrial operations. Calculating the optimal sequence for carrying out tasks therefore makes it possible to increase process productivity.
A robotic system has been developed at the University that identifies the parts placed in its working environment and can calculate the optimal operating sequence to carry out the task in the shortest possible time.
Our solution to the problem has been to develop a set of algorithms based on decision trees, mathematical models of binary integer programming and reinforcement learning to calculate the best sequence for handling the parts.
TECHNOLOGY APPLICATIONS
Robotic handling optimisation can be applied in the following areas:
- Port Logistics and Warehousing: Optimised management of the movement of containers or pallets in confined spaces.
- Injection Industry: Manufacture of multi-component products in which precision and assembly order are critical.
- Fashion and Footwear: Sorting and grouping cut components for subsequent joining on a production line.
- Urban Mobility: Although focused on industry, the algorithmic approach can be applied to optimising shared vehicle fleets to minimise fuel consumption.
ADVANTAGES AND INNOVATIVE ASPECTS
This technology is already applied to the optimisation of freight transport and to taxi services. In our case, we apply similar techniques to robotic manufacturing processes, an area in which other universities or research centres do not specialise.
Applying these techniques to robotics introduces additional complexity because the probability of collisions between the transport elements (robotic arms) is much higher and more restrictive than in the transport of goods or passengers.
DEVELOPMENT STAGE
The system is currently at TRL 4 (validation in laboratory environments). Work is currently under way to implement the system in a simplified industrial environment. The technology is expected to reach TRL 5 (validation in a relevant environment) by the end of 2022.
APPLICATION SECTORS
- Industrial Automation and Robotics.
- Logistics and Freight Transport.
- Textiles, Footwear and Components.
- Machinery Manufacturing and Automotive Industry.
PRINCIPAL RESEARCHER
Carlos Pérez Vidal – University Institute for Engineering Research (I3E)