Robocoast Challenge: Measuring with AI
Robocoast is launching a challenge related to production equipment measurement data, where deviations are sought using artificial intelligence. The best solution proposal will be piloted. The company providing the challenge is a major metal industry company from Satakunta.
The challenge provider’s production equipment provides measurement data that varies under different circumstances. However, even though measurement data is always similar under the same circumstances, deviating values can also be found. Artificial intelligence is required to detect these values. Artificial intelligence capable of detecting deviations can be used in various production settings.
A large amount of data is collected from production machines and analysed using various analysis tools. However, artificial intelligence is not yet available for this purpose, even though data is available for machine learning. The aim of the challenge is to launch a pilot, which will be used to:
- Understand what is required in order to teach artificial intelligence, and how much work is required to teach new circumstances to AI
- Learn what kind of technical skills the use of artificial intelligence requires
- Find out what kind of requirements artificial intelligence imposes on IT infrastructure
Evaluation data, including both normal data and random deviations, will be sent to the solution providers later on in an Excel/CVS file. Because the pilot will not be connected to a real machine or system, the production equipment will be simulated, for example, by sending real-time measurement data to a machine learning pilot at regular intervals (e.g. one measurement per second).
The objectives of the pilot:
The machine learning model will be taught a “normal situation” as well as two or three exceptional cases. The average value of the measurement results may change, so the trendline of the results must be taken into account.
- Some kind of notification or alert should be given when deviations occur.
- The solution does not need to communicate with any other system.
- The measurement data must be provided in Excel/CVS files.
The challenge is open to everyone. Participants will not receive any compensation for taking part in the challenge. However, the winning company may receive a maximum of 10,000 euros for creating a demo based on their proposal. The participants will have access to the data required for planning a demo.
Questions regarding the challenge:
Questions regarding the challenge can be submitted by email to:
The solution proposal must be submitted in writing as a free-form document to email@example.com by June 9, 2019.
The proposal must include:
- Contact information for the company/companies proposing the solution
- A description/presentation of the solution proposal, including observations and results
- A (non-binding) indicative cost estimate or other proposal for follow-up measures
Feasibility and applicability will be emphasized when assessing the solution proposal. The solution proposals will be assessed by the company providing the challenge.
All solution proposals will be handled confidentially. The intellectual property rights of the solution will remain with the solution provider until the prize money is paid, after which time the code will be become public. The winner has the right to refuse the prize, in which case the intellectual property rights will remain with the solution provider. In that case, the solution provider and the company providing the challenge may freely agree on further action. The intellectual property rights of participants are protected by means of a contract between Robocoast and the participants stating that only the prize-winning code will be published.
The challenge will be organised as part of the “Robocoast R&D Center – Robotics Living Lab for companies” project funded by the European Regional Development Fund and the Regional Council of Satakunta.