Company cases
RoboAI company projects combine the needs of the companies, the expertise of the professionals and the fresh and innovative views of the students.
Company projects are an important part of RoboAI activities. Research teams produce and test new research data with projects to be used by the companies. Academy students get to learn, innovate and try new ideas in company projects in a practical manner.
This page shows our latest and most significant projects.
Theses
Student Jenni Alatalo from Satakunta University of Applied Sciences created an automation plan for the packaging line of a bakery. The thesis was commissioned by Pekan Parhaat bakery and it was implemented as a part of 5VTA project (Five effective technology steps in food industry SMEs project). The project partners include Satakunta University of Applied Sciences, Seinäjoki University of Applied Sciences, Regional Council of Satakunta and Regional Council of South Ostrobothnia.
The purpose of the thesis was to find out how the cutting and packaging of swiss roll slices could be automated in a bakery enterprise. The most appropriate devices were found based on the production volume, the costs were estimated, and a simulation was made according to the most cost-effective automation plan. The goal of the thesis was also that automation would enable more effective use of time, help in gaining savings based on accelerated packing, and increasing the production volume.
When selecting the devices, besides the price special attention was also paid to the suitability of the devices in a bakery environment, where hygiene is extremely important. The space required by the devices is also a significant factor for the working to be pleasant and safe. A major benefit in the selection of the devices is the possibility to use the devices in the manufacturing process of other products, thus making the investment more beneficial.
Three plans of a different level were drawn up for automation:
- In the first plan, all the stages related to cutting and packaging are automated by utilizing automated devices provided for bakeries. The only manual labour is moving the packed products to a cardboard case.
- In the second plan, a collaborative robot cuts the pieces by a special tool designed for it. The same robot moves the cut products to individual packages that are conveyed through flow pack packing machine. From the packing machine the products continue to a rotating table, which an employee empties at appropriate intervals to cases.
- In the third plan, only the flow pack packing machine and the following rotating table are utilized. Thus, the cutting, packing of individual products and packing of plastic coated products into cases is carried out by manual labour.
When cost calculations were made based on these plans, it was discovered that the cost of the automation with the first plan would be more than 200 000 € without the installation and programming costs. The devices in the second plan cost altogether approximately 70 000 € and in the third plan circa 40 000 €. Based on these plans, the most cost-effective solution in this situation was plan number two.
Read more about Jenni Alatalo´s thesis
Robotics Academy company projects
Robotics Academy was commissioned by Boliden Harjavalta nickel smelting plant to utilize machine vision in their process. The possibilities of machine vision were studied by students. The aim was to deploy machine vision to study the concentrations of different constituents in slag after the smelting process.
A system where computer vision is applied for industrial purposes is called machine vision. The system consists of a light source, a target to be imaged, a camera, a computer with an image processing software that automatically interprets the image.
Machine vision systems mainly execute tasks that are strictly pre-programmed. In this commission concentrations given by the customer were recognized on a conveyor belt. The used application was Cognex application In-Sight. In a smart camera all the image processing and calculations happen in the camera itself.
During the project the students learn how to utilize different camera and light options. As a whole, the project gave mutual benefit. The competence of the students grew and the commissioner received the information they hoped for. With the knowledge gained, projects requiring machine vision can be met better.
Robotics Academy was commissioned by Oras Group to research the suitability of a robot for assembly work. The goal was to automatize a five-part assembly line utilizing an assembly robot.
The robot chosen for the project was a dual-arm YuMi robot, which has been developed to assemble small parts. The robot capable of real cooperation has arms that change position flexibly, part location function based on camera and a precise control system.
In the project, the tasks of Robotics Academy students included the programming and simulation of the robot, utilization of machine vision and the design and implementation of the parts connected to the robot, i.e. jigs, pallets, and layers, by 3D-printing. The project gave the students a possibility to comprehensively learn the functioning and limitations of the robot. The primary goal was reached, i.e. the assembly task was executed by the robot.
“Cooperation with Robotics Academy was easy and we were on the same wavelength already at the first meeting. The assembly of our product varies and with this project we definitely became aware of potential new ways of functioning. The results were interesting and they are a good basis for further consideration. I believe both cooperation parties gained benefit from the project.”
Maarit Ruohola
Oras Group
Robotics Academy was commissioned to assemble certain magnets to a required length and form. The possibilities to use an industrial robot for this purpose were studied by students. The goal was to research whether this given task could be implemented by using different industrial robots.
An industrial robot is a computer-controlled machine that can execute the same repetitive task with the same accuracy and speed several times. An industrial robot is a general-purpose machine, and the same robot can execute several different tasks, depending on the program and the purpose.
The commissioner was Neorem Magnets, the only magnet factory in Finland. Their main products are different magnets and magnetic solutions, and they produce high-quality magnets for the different fields of industry utilizing state-of-the-art technology in the field.
Two robots were used in the project: UR-5 and ABB IRB-120. The robot chosen to finalize the project was UR-5 because it met the requirements and hopes of the commissioner better. UR-5 is easy to program and sufficiently fast to execute the task in question.
In the project the students learned how to program and utilize the different features of the robots in an industrial environment, as a whole the benefit gained was mutual. The students´ competence grew remarkably, considering that this project was the first touch to the wonderful world of robotics to some students. With the competence gained, future robotics challenges can be met even more efficiently.
Robotics Academy got a challenge from Neorem Magnets connected to measuring the dimensions of the magnets. The dimensions of the polygon-shaped magnets should be measured with the accuracy of hundredths of a millimeter.
The challenge was tackled by first conducting a project, where Belgian and Spanish exchange students studied how the measuring of magnet dimensions could be executed by a smart camera. Already at the beginning of the first project it was known that even the best of smart cameras with the highest of resolutions were not accurate enough for the accuracy of hundredths of a millimeter. The goal was set to find out what kind of tools and setup were needed for the measuring to succeed.
At the beginning of the project Colin and Nicolas from Belgium and Jaime from Spain got acquainted with the analysis program of the smart camera and built a setup to be able to capture an image of the magnet as precisely as possible, in stable conditions and repeatedly. When the basic functions of the analysis program had been studied, they started to execute the measuring program. With the measuring program, machine vision system recognizes a piece, finds its edges and measures the distances between the edges in relation to each other, and measures the angles between the edges. When the program was ready, the students tested imaging and measurements by measuring all the pieces three times on both sides. Thus, data was received on the repeatability and measurement error of the analysis program.
Based on this first project, the results showed that measuring of magnet dimensions was possible by machine vision, and by measuring technique the repeatability and speed of the measuring process could be improved. Based on this project, the goals of the next project were defined, i.e. building more precise imaging arrangements and an applicable analysis program. Moreover, designing a more accurate image jig was set as an objective, as it helps to stabilize the magnet on the imaging platform.
Implementation of the first project was a very pleasant project work for the exchange students. They learned how to program smart camera analysis program in an environment previously unknown to them. They also genuinely focused on presenting the results concerning different pieces and measurements.
“The students had created a really impressive program in a short time and a lot of data was compiled for the report, which is particularly valuable when evaluating the functionality of the program and analyzing the areas for improvement. With cooperation we will receive valuable information on different machine vision systems and their price categories and applicability for our different needs to support the investment decision.”
Jukka Hissa
Production development engineer
Neorem Magnets
Company projects supported by funding
Problem/Need
More and more people meet challenges of wellbeing of the mind in their everyday life. For example fear of social situations, performance anxiety and stress are commonplace. Especially young people need new solutions that suit their world to support the traditional paper and pen–methods. Technology offers much needed low threshold instruments to work on these challenges.
Cost-effective and easily available solutions to promote the wellbeing of the mind
Exposure
Exposure therapy is one of the most efficient methods in treating anxiety. According to studies, 60 – 75 % of people treated with exposure treatment feel they get some sort of relief in their condition and the effects are long-lasting. Exposure therapy is based on facing the fear. When persons are exposed to the targets of their fear long enough, the mind adapts to the stimulus that causes fear, and thus it will not cause a stress condition anymore. Virtual reality is a potential tool for exposure therapy because it helps to create experiences that feel real in a safe environment. Digital environment also assists in making all kinds of changes (e.g. adjusting the number of stimuli and the progress of the situation), which in the real world would be cumbersome or expensive to implement (cf. fear of flying and flying).
Example: Exposure application for agoraphobia
In the demo the user is at a store, and the goal is to queue and pay for the shopping. The demo uses a real 360 degree-image that is watched with virtual reality headset. In the demo the user can proceed from one interaction (go to the queue) to another (do business with the salesperson) at their own speed. The aim of the exercise is to result in an anxiety response, wait for the stress condition to ease and then continue with one step at a time, at own speed.
Relaxation
VR technology can also be used to support relaxation. The user can be taken virtually to diverse relaxing environments (e.g. the nature). Therapeutic tools can be added to the virtual world, which help in boosting relaxation and recovery. For example breathing exercises added to the virtual world can help to activate the parasympathetic nervous system to promote recovery.
Example: Breathing exercise
The relaxation demo uses 360 degree-image, which has been filmed in the fell scenery in Lapland. Controlling the relaxation exercise takes place by a figure created for the demo. In addition, a ball that becomes larger and smaller at the right rhythm is utilized in the demo to guide the breathing to boost relaxation and recovery.
“At its best, virtual reality offers completely new possibilities of implementing e.g. exposure therapies. Utilization and continuous development of virtual reality and various technological applications is reality and the role of technology in psychiatric nursing will most likely increase even more in the future.”
Anna Mäkelä
Adolescent psychiatrist, locum chief psychiatrist, Satakunta Hospital District, Adolescent Psychiatry Outpatient Clinic
The field of rehabilitation needs novel means to support independent rehabilitation. However, the customers are a heterogeneous group with individual needs for rehabilitation and with different level of functional ability. The games utilize many elements that help to increase motivation, enthusiasm and commitment. So-called “beneficial” elements can be embedded in games, and then they are called serious games. The challenge in serious games is adapting the contents of the game for different user groups. Especially a solution that is modular and customizable according to the user is needed.
The idea of the game is to activate the user for a light physical exercise, and at the same time practice coordination and perception. With those seriously disabled, the main goal is participation and enabling new experiences.
Controls
To increase physical activity, body movements are utilized to control the game. The game is controlled by a small sensor that monitors the position (inclination sensor). Sensor data is sent by Bluetooth to a mobile device, which has the game installed. The sensor can be attached to different limbs or to different items that the player moves. Thus the control movement needed can be modified according to the user´s needs. Modularity can be added by customizable control solutions that are 3D printed. 3D printing makes it possible to manufacture entirely individual controls when the sensor is connected to the control.
FDM (Fused Deposition Modeling) was selected as 3D printing technology. A plastic thread is fed through a nozzle and heated to a temperature suitable for the material. The molten plastic is extruded onto a build platform, layer after layer and finally forming the entire product. The used printing material was PLA (polyactide) because of its printing properties, such as low printing temperature and dimensional accuracy.
Here the example controls used were:
- a balance board when the movements can be made by using feet (by sitting on a chair or standing on the balance board)
- hand controls, with various ways of gripping (the handles can be changed and the control can be attached to the body)
- head control, which enables controlling without the use of hands or feet
Games
As an example, a simple labyrinth game was implemented where the player (a mouse) needs to collect hearts (picture). The game can be adjusted according to different users. For example, the size and speed of the game elements, the amount and location of the mazes and the number/quality of the opponents can easily be modified according to the needs of the user. It is essential to adjust the level of difficulty to the player´s abilities. The game was realized to be adaptive – it adapts to the player´s abilities by raising the level of difficulty with opponents and mazes. The game begins as simply as possibly without mazes or opponents. Then e.g. those with challenges in perceptual ability or motoric skills can concentrate only on collecting their first heart during the whole game, without the fear of the game being too difficult and thus killing the motivation. On the other hand, an advanced player does not get bored because the game becomes automatically more difficult, when more hearts are being collected. The same control technology can be utilized with various games, and they are being worked on all the time.
"It is important that rehabilitation is also nice. The games are a good aid to make rehabilitation more meaningful. We have taken these new tools both into individual practice and as a part of group practice"
Helena Myllymäki
Quality Manager, Occupational Physiotherapist, Rehabilitation Centre Kankaapää
Tampere University Pori Unit does also a lot of company cooperation as commissioned cases. Currently, a reliable analysis mechanism for slaughterhouse waste is being developed for Honkajoki Oy. The analysis mechanism based on hyperspectral imaging could be used in the further processing of slaughterhouse waste.
The unit had a Specim Ltd hyperspectral camera in use, by the help of which the spoilage level of slaughterhouse waste could be recognized as well as foreign objects, such as metal, plastic, glass and the water content of slaughterhouse waste. Laboratory samples taken of slaughterhouse waste are used as reference values for hyperspectrum data. The images have been taken together in cooperation with the researchers of University of Tampere and Specim Ltd.
The measurements of hyperspectral camera are fed to the Computing Unit of Tampere University neural network (artificial intelligence), where the neural network strives for learning the causal relation between hyperspectral camera measurements and laboratory measurements taken from slaughterhouse waste. Teaching neural network requires several interim stages because there is plenty of hyperspectral data and before neural network can be taught, the amount of data needs to be reduced without, however, losing any essential data. Thereafter, the size of the neural network must be clarified and optimum web learning parameters found before the actual teaching of the neural network can take place. Moreover, Dyme Solutions Oy has implemented a data acquisition solution with which data from Honkajoki Oy´s different processes is gathered in one place. Researchers from Tampere University try to find cause-and-effect relations in the collected data, which would lead to development measures in different processes, e.g. review of energy consumption and reduction of energy consumption in future or predicting the breakdown of components related to process.
This example presents the planning and implementation of machine visions system that guides the robot to pick chocolate slabs on the line and pack chocolate slabs of different flavor in their own cases.
The machine vision system was implemented as a traditional machine vision system consisting of a machine vision camera, a LED light that lights the chocolate slabs and an analysis software running in the computer. The machine vision camera always takes an image of the chocolate slab at the same point of the conveyor. ( Picture 1. Chocolate slabs being inspected by machine vision system)
The analysis software running in the computer recognizes the flavour marking on the chocolate slab as well as the location of the chocolate slab, i.e. its coordinates on the conveyor (picture 2). When the robot receives the information of the coordinates, it picks the chocolate slab from the conveyor and packs it into the right case based on flavour information.
Picture 2. Chocolate slabs recognized by the machine vision system and their coordinates
Five effective technology steps in food industry SMEs – 5VTA is a projected implemented by Satakunta University of Applied Sciences and Seinäjoki University of Applied Sciences, which has been given funding by Regional Council of Satakunta and Regional Council of South Ostrobothnia from European Regional Development Fund.
When we want a mobile robot to convey food packages to be boxed or boxes to be palletized, we must be able to call the mobile robot to convey the products if needed. In 5VTA project, a call application for mobile robots was made by way of example.
With the call application, a worker can call the robot to go and get the products, for example from the packing machine to be palletized. The call application (picture 1) functions on a mobile device and the worker can choose where the mobile robot is called and where it is sent to when the products are onboard. The call application enables also making a reservation for the robot, i.e. if the robot is busy, it can be reserved to come next to a determined place. The application can also be used to create a route for the robot according to which it can proceed automatically from one operating point to another, or first to the packing department and then to the dispatch department.
Picture: Sending the robot takes place by choosing the target, where the robot needs to convey the products.
Five effective technology steps in food industry SMEs – 5VTA is a projected implemented by Satakunta University of Applied Sciences and Seinäjoki University of Applied Sciences, which has been given funding by Regional Council of Satakunta and Regional Council of South Ostrobothnia from European Regional Development Fund.
In this example different bakery products are handled by a robot gripper, whose grip force can be adjusted so precisely that the products that are easily deformable do not stretch or get squashed.
The bakery products in this example are a doughnut and a mini-doughnut. The light touch gripper is used in this example to:
- Handle the doughnuts by lifting them on the inner edge
- Handle the mini-doughnuts by lifting them on the outer edge
Light touch grippers like these enable flexible manufacturing as they can be used to handle products of all forms, without affecting their shape. Thus, several different products can be simultaneously handled on the production line.
The locations of doughnuts and mini-doughnuts are recognized by PickIt 3D camera that guides the robot to pick the bakery products from the right location.
Five effective technology steps in food industry SMEs – 5VTA is a projected implemented by Satakunta University of Applied Sciences and Seinäjoki University of Applied Sciences, which has been given funding by Regional Council of Satakunta and Regional Council of South Ostrobothnia from European Regional Development Fund.
Do you wish to outsource a project or get a new production line simulated?
Contact us!
Petteri Pulkkinen
Research Director
tel. 044 710 3296
petteri.pulkkinen@samk.fi
Subscribe to RoboAI newsletter
Get the latest news and updates directly to your inbox.