Co-operation cases
RoboAI co-operation cases combine the needs of the companies, the expertise of the professionals and the fresh and innovative views of the students.
Co-operation cases are an important part of RoboAI activities. Experts 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 the cases in a practical manner. This page shows our latest and most significant cases.
RoboAI Industry
In a Proof of Concept pilot carried out with Pori Energia, a computer vision and AI-based solution was developed to automatically detect whether employees are wearing the required personal protective equipment (PPE), including safety helmets, safety glasses, hearing protection, and high-visibility vests. The aim was to explore how artificial intelligence could enhance workplace safety and help reduce risks caused by missing protective equipment.
The pilot system identified workers' safety equipment in real time and displayed a visual alert whenever mandatory protective gear was missing. The solution was also demonstrated at Pori Energia's Safety Day, where it showcased the practical potential of AI-assisted safety monitoring in industrial environments.
The pilot provided a strong foundation for further development and demonstrated that AI can play an important role in creating safer workplaces and improving the efficiency of occupational safety monitoring.
In collaboration with Biolan Oy, an initial experiment followed by a Proof of Concept (PoC) pilot explored how palletizing products of different sizes could be automated using a movable collaborative robot cell. The objective was to develop a flexible solution that could be deployed across multiple production lines while reducing physically demanding manual work.
The project included the design of the robot cell layout, the development of a prototype gripper for handling products of varying sizes, and the evaluation of safety solutions. Based on the successful preliminary trials, the PoC pilot validated the concept using a real collaborative robot and actual products, providing valuable insights for further development.
The results demonstrated that collaborative robotics is a promising solution for flexible palletizing in industrial environments. The pilot also provided Biolan with practical recommendations to support future implementation and automation planning.
Read more about the experiment and the pilot (in Finnish).
In collaboration with Nurmilo Oy, an initial experiment followed by a PoC pilot explored how the picking of small metallic and shiny parts could be automated using machine vision and a collaborative robot. The aim was to find a reliable solution for a repetitive task requiring precision.
The tests compared different machine vision methods and showed that backlighting significantly improves the detection of shiny parts. In the PoC pilot, the solution was developed further into a compact robot cell, where a UR5e collaborative robot and a Cognex smart camera picked parts in a continuous cycle.
The pilot demonstrated that the solution can be implemented with the equipment setup preferred by the company. Nurmilo Oy also gained concrete data on the solution’s reliability, cycle time, and potential for adapting the flexible robot cell to future production needs.
Read more about the experiment and the pilot (in Finnish).
In collaboration with Piir-Group Oy, an initial experiment followed by a Proof of Concept (PoC) pilot explored how furniture panels of varying sizes could be automatically picked, handled, and palletized using a collaborative robot. The goal was to develop a flexible solution capable of identifying and handling different panels with minimal manual intervention.
The project combined QR code identification, machine vision, and collaborative robotics to recognize panel dimensions and optimize handling. The concept was validated in a PoC pilot using a real collaborative robot, while also evaluating key technical requirements such as the gripper design, camera setup, and robot reach.
The pilot demonstrated that the concept is well suited for automating furniture manufacturing processes. It also highlighted how product identification data can support wider production digitalization, including warehouse management, traceability, and production monitoring.
Read more about the experiment and the pilot (in Finnish).
In a Proof of Concept pilot carried out with Neorem Magnets Oy, a machine vision system was developed for the precise measurement of small production components as part of the company's quality assurance process. The objective was to improve measurement accuracy, accelerate inspection, and evaluate the suitability of the technology for industrial use.
The solution was developed jointly using a telecentric lens, an industrial machine vision camera, and MVTec HALCON software. The measurement algorithms were refined to better reflect real production conditions, while the company also gained practical experience in developing and utilizing machine vision applications.
The pilot resulted in a more accurate and reliable measurement system that improved inspection efficiency and provided a solid foundation for expanding the use of machine vision in production.
RoboAI Health
Dyme Solution Oy's Florie pressure sensor was tested in the spring of 2023 in the Naskali residential unit of the Developmental Disability Services of the Satasairaala hospital.
The aim of the test was to evaluate how the sensors placed at the foot of the bed enable real-time monitoring of movement and its impact on the night-time safety of the residents and the night-time peace of mind of the entire unit. The functionality and usability of the Florie sensors as a product and the user-friendliness and functionality of the user interface were also evaluated.
In the test, sensors were installed under the feet of the bed of a resident in the Naskali housing unit, which, through changes in weight, detect when the resident leaves the bed and send the information to Dyme Solution Oy's cloud service. Carers were alerted to the bed exit by an app downloaded to their mobile phones.
The product was found to be particularly suitable for residents with poor balance who are at risk of falling, "poor footed" but who are alert or restless at night or may struggle to get up despite being on the edge of the bed. The product is useful when the unit is large, with long distances to cover, a lot to do at night and few staff to supervise.
Henna Leppänen's Master's thesis on Florie (Theseus, in English)
Research centres as business partners - RoboAI and Dyme Solutions Ltd (Youtube, in Finnish)
More and more people are facing mental well-being challenges in their everyday lives. For example, fear of social situations, performance anxiety and stress are common for many. For young people in particular, new solutions are needed to support traditional pen and paper methods. Technology offers much-needed low-threshold tools to work through these challenges.
Cost-effective and accessible solutions to promote mental well-being.
Exposure
Exposure therapy is one of the most effective methods of treating anxiety. Studies show that 60-75% of people treated with exposure therapy experience some relief and the effects are long-lasting. Exposure therapy is based on confronting fear. When a person is exposed to the object of their fear for long enough, the mind adapts to the fearful stimulus, so that it no longer causes stress. Virtual reality is a potential tool for exposure therapy because it can be used to create authentic experiences in a safe environment. The digital environment also allows for a wide range of modifications (e.g. adjusting the amount of stimuli and the progression of the situation) that are difficult or costly to implement in the real world (cf. e.g. fear of flying and flying).
Example: Exposure app for fear of public places
In the demo, the user shops in a store. His goal is to queue and pay for his purchases. The demo uses a real 360 image viewed through virtual glasses. In the demo, the user can progress from one interaction (go to the queue) to another (deal with the salesperson) at his/her own pace. The purpose of the exercise is to elicit an anxiety response, wait for the stress to subside and then proceed one step at a time, at your own pace.
Relaxing
VR can also be used to support relaxation. The user can be virtually transported to a variety of relaxing environments (e.g. nature). Various therapeutic tools can be added to the virtual world to enhance relaxation and recovery. For example, breathing exercises in the virtual world can be used to activate the parasympathetic nervous system to promote recovery.
Example: breathing exercise
The relaxation demo uses an authentic 360 image shot in a mountain landscape in Lapland. The relaxation exercise is guided by a character built into the demo. In addition, the demo uses a ball that increases and decreases in rhythm to control breathing to enhance relaxation and recovery.
"At its best, virtual reality offers a whole new range of possibilities, including exposure therapies. The use and continuous development of virtual reality and various technological applications is a reality of today and the role of technology in psychiatric care is only likely to grow in the future."
Anna Mäkelä
Specialist in Adolescent Psychiatry, Senior Physician, Satakunta Hospital District, Pori Adolescent Psychiatry Clinic
The rehabilitation field needs new ways to support independent rehabilitation. However, the clientele is very heterogeneous, with individualised functional capacity and rehabilitation needs. Games make use of many elements to increase motivation, enthusiasm and engagement. Games can also embed 'benefit' elements, and are typically referred to as 'benefit games'. The challenge with utility games is to adapt the content of the game to different user groups. In particular, there is a need for a modular solution that can be adapted to the user.
The idea of the game is to engage the user in light physical exercise while at the same time training coordination and perception. For the most severely disabled, the key objective is to engage and enable new experiences.
Controllers
To increase physical activity, body movements are used to control the game. The game is controlled by a small position sensor (tilt sensor). The sensor data is sent via Bluetooth to the mobile device where the game is installed. The sensor can be attached to different limbs or different objects that the player moves. In this way, the required steering movement can be modified according to the user's needs. Modularity can be increased through customisable controller solutions that are 3d printed. 3d printing allows for the production of fully customised controllers, as long as the sensor can be attached to the controller.
The 3D printing technology chosen is FDM (fused deposition modelling), where a plastic wire fed into a nozzle is heated to a temperature suitable for the material. The molten plastic is then extruded onto the printing substrate in layer after layer, eventually forming a complete part. PLA (polyactide) was used as the printing material because of its printing properties, such as low printing temperature and its dimensional accuracy.
Here, the following were used as example guides:
- a balance board, allowing movements to be made with the feet (sitting on a chair or standing
on the balance board)
- hand controllers with multiple ways of gripping (handles can be interchanged and the control can be attached to the body)
- with a head control, allowing you to steer without using your hands and feet
Games
As an example game, a simple maze game was implemented where the player (mouse) is supposed to collect hearts (picture). The game can be adjusted for different users. For example, the size and speed of the game elements, the number and location of mazes, and the number/quality of opponents can be easily changed according to the user's needs. The key is to match the difficulty level with the player's abilities. The game was implemented to be adaptive - it adapts to the player's abilities by increasing the difficulty level through the use of opponents and mazes. However, the game starts as simply as possible without mazes and opponents. This allows people with perceptual and motor challenges, for example, to concentrate on collecting the first heart of their game, without the game's difficulty killing their motivation. On the other hand, the advanced player will not get bored, as the game automatically becomes more difficult as more hearts are collected. The same control technology can be used with a wide range of different games and is currently being worked on.
"It is important that rehabilitation is also fun. Games are a good tool for making it more meaningful. We have introduced new tools both for individual training and as part of groups."
Helena Myllymäki
Quality Manager and Occupational Physiotherapist, Kankaanpää Rehabilitation Centre
In cooperation with RoboAI Health and Rauma Fysikaalinen Hoitola Oy, a virtual reality-based rehabilitation game Mokke (Goalie) was developed. The development was driven by the leading idea of adjustable features according to the client's individual goals and the use of motion data during gameplay. An essential part of the game's suitability for therapeutic use is the ability to control the game's settings externally.
The need to develop the game arose from the owners of Rauma Physical Therapy Ltd's (RFH) idea to introduce a virtual game that activates the whole body and especially the lower limbs. The Mokke game developed at SAMK is controlled by the upper limbs, but the lower limbs can be activated as required or the exercise can be done sitting down. The RFH physiotherapists wanted the game to be simple but challenging for the client and individually adjustable to maintain motivation for the game and ensure that the training was goal-oriented. Monitoring recovery and goal achievement requires measuring and observing the client's activities. It is therefore important for physiotherapists to be able to see the data on range of motion, range of motion, speed and accuracy that is accumulated through play.
The use of virtuality in rehabilitation, especially "immersive virtuality" using VR glasses, reduces pain perception (Wong, Tse & Qin, 2022). In the Mokke game, the training aims to activate the client to move and perform exercises with natural movement pathways with less pain perception. During the game, the trainee may perform larger ranges of motion and more repetitions than in a traditional exercise format.
A personalised and fun way to rehabilitate 
In the Mokke game, the rehabilitee acts as a football goalkeeper using virtual glasses (Figure 1). A key feature of the game is the ability to adjust game settings using a remote control app (Figure 2). For example, by adjusting ball arrival times, positions, speeds and ball size, the settings can be adjusted from a physiotherapy perspective to suit the needs and goals of the rehabilitee.
The game itself collects data generated by the virtual laser controllers for analysis by the physiotherapist. Data is collected on the position of the controllers, the speed and acceleration of movement. The remote control application displays the maximum and minimum values of movement frequencies, maximum and minimum speeds, exercise duration and the number of balls intercepted in the coordinate system. In addition, visual descriptions of the position and speed data of the controllers are provided. The exercise data collected in the database will be used to monitor the progress of the exerciser and to plan new exercises.
The training area must be large enough to use the full capacity of the game. A football goal in this game is therefore equivalent in size to a real football goal. However, it is possible to play in a smaller area, so that the regulations can be adapted to the needs of different clients. Further development of the game is possible and the use of lower limbs as controls would be an interesting variation. At this stage it was not yet possible to attach the controllers to the lower limbs, but different sensors or attachment mechanisms could allow this.
RFH physiotherapists Hannakaisa Mäkitalo and Hanna Teerialho think that the Mokke game is an interesting addition to rehabilitation, as is the whole world of VR in general. The game could also be used in other sports, such as volleyball or ice hockey. It would be great if it were possible to choose between different sports in the same simulation."
The Mokke game has been developed in a project funded by the Ministry of Education and Culture (Special grant for the development of competence and network development in higher education in early childhood education and care and education personnel) called Network cooperation to boost the accessibility of AI-based virtual technologies in rehabilitation.
The Mokke game has also been tested as an aid to goalkeeper coaching at Pallo-Iirot Football club. Read more about this on the Satakunta Testbed website!
Further information:
Taina Jyräkoski
taina.jyrakoski@samk.fi
SAMK implemented the testing of the Vitacam application in May-August 2024 in cooperation with Diakon Services. The testing provided valuable information on the usability and usefulness of remote measurement of vital functions in the context of home care. Vitacam is a medical software device that measures vital functions (respiratory rate, pulse, heart rhythm) using a web or mobile camera. The aim of the testing was to assess the application's suitability for monitoring early changes in the well-being of elderly people.
Implementation of the test
12 independently living elderly people and two nurses participated in the test. The testers used the app for two months and took measurements at least three times a week. The nurses received weekly reports of the measurement results and responded to any abnormal values as necessary.
Key findings
- Testers found the app easy to use and useful.
- Measurement success rates were high (pulse 99%, respiratory rate 95%).
- Technical challenges were mainly related to updates of the app and the use of phone cradles.
- Testers needed clearer instructions and more information on the measurement values.
Conclusions
The Vitacam app is well suited for stand-alone monitoring of vital signs. Further development is recommended to improve the instructions for use, to make the monitoring of measurement results more intuitive and to develop additional guidance on measurement conditions. These recommendations have already been taken into account in the further development of the application.
You can read more about the case on the Satakunta Testbed website (In Finnish)
In January-February 2024, the IKI project of Satakunta University of Applied Sciences tested Tactic's games with elderly people in Diakonkylä apartments. The purpose was to find out how the games could be developed to suit this target group. Kimble, Kimble Junior, Find and Find! Nature Trip and the Ötökät memory game were tested in five game events. There were a total of 56 participants, and they played in groups of 3–4 people. Project workers supervised the games and collected feedback. Some of the players required operational management during the game.
Key findings:
- Visual perception: In some of the games the font size of the instructions was too small and the score sheets were difficult to read. The white sand on the hourglass was hard to see, suggest black sand. On the board, green, blue and purple were mixed and the lighting reflected distractingly off the boards.
- Cognition: some images were unclear or difficult to perceive, and naming the images could help players. In the memory game, the number of cards divided opinion - the instructions could include a recommendation to start with a smaller deck. The rules of the games should have different levels of difficulty.
- Physical activity: the cards were mostly easy to handle, but were too small for people with motor challenges. The game boards moved on the table, and sliding barriers could be added underneath them. The buttons were too small for many people and it was difficult to turn the hourglass.
- Social situation: playing together was perceived as socially rewarding, but the possibility to play alone should also be mentioned in the guidelines.
The testing shows that Tactic games have potential to engage older people, but clearer instructions, stronger visual contrasts and more physically accessible components are suggested to improve accessibility and usability.
You can read more about the case on the Satakunta Testbed website. (in Finnish)
The testing of the Neppie application was carried out from August to December 2024 under the coordination of Satakunta University of Applied Sciences. The testing was carried out in the Tukena Woimaamo, where the application was used as part of the daily routines of the clients. The product under test was the Neppie application developed by FunctionApp Oy, designed to support the management of everyday routines.
The aim of the testing was to assess how Neppie supports the activity management of clients and the work of professionals and to collect user experiences for the development of the application. In addition, the aim was to map out what special features are associated with application testing among Woimaamo's client group.
Test implementation
Seven clients and one supervisor from Woimaamo participated in the test. The app was mainly tested in cleaning tasks. During the test period, the app was used with two tablets twice a week, and the task sequences contained 4-8 steps. At the beginning of the test, the staff was trained to use the app. Feedback was collected through an anonymous questionnaire.
Findings and results
- Neppie supported clients in managing their activities and increased their self-direction.
- Staff suggested for smoother creation of chains.
- Scalability did not work optimally.
- The application is best suited for simple and repetitive tasks.
- Gamification was not considered necessary for adult users.
- Feedback found the app to be clear and user-friendly.
- Application testing should allow for flexibility in the testing setup and the ability to react to changing situations, as well as clear instructions for partcipants.
Suggestions for improvement
- Faster and more flexible task chain creation.
- Solving the problem of image scaling.
- Ability to hide tasks and edit chains on the fly.
- Visualisation of task chain progress and "break step" feature.
Summary
Neppie is a powerful tool for activity management, but improving its chain management would increase its usefulness and usability even further. Customers found the application easy to use and a clear tool for everyday life.
You can read more about the case on the Satakunta Testbed website. (In Finnish)
RoboAI Cyber
Collaboration Accelerated the Development of Device Testing
As part of the Kytee project, Polarteknik Oy participated in a device testing pilot at the RoboAI Cyberlab to evaluate the cybersecurity of its electronic door controller and to develop a testing approach that could also support future product assessments. The pilot provided valuable insights for both the company and RoboAI Cyberlab, demonstrating the importance of close collaboration in advancing cybersecurity.
During the pilot, the focus was not only on assessing the device's cybersecurity but also on evaluating how the testing process itself could best meet the needs of companies. For Polarteknik, obtaining an independent assessment of its product's security was essential, as was developing a repeatable testing model to support product development and meet increasing customer and regulatory requirements.
One of the greatest benefits of the pilot was the opportunity to gain new perspectives, learn together, and improve testing practices collaboratively. The testing confirmed several of the company's own observations while also identifying new areas for improvement that might not have been discovered without external expertise. At the same time, the pilot generated valuable knowledge for further developing device testing services to better serve industrial needs.
From RoboAI Cyberlab's perspective, the pilot marked an important step in building its device testing services. Working closely with an industrial partner provided a deeper understanding of companies' requirements and helped refine both the testing environment and the testing methodology to better address future cybersecurity challenges.
Both Polarteknik and RoboAI Cyberlab view cybersecurity as a continuous process of improvement in which trust, open collaboration, and knowledge sharing play a central role. Emerging regulations such as the Cyber Resilience Act (CRA) further emphasize the need for systematic and reliable methods for demonstrating product cybersecurity.
The pilot demonstrated that device testing is much more than evaluating a single product. At its best, it is a collaborative process that helps create more secure products, strengthens companies' cybersecurity capabilities, and develops testing practices that benefit the wider industrial ecosystem.
Making cybersecurity part of everyday operations
Knobbi Oy participated in a pilot of the cybersecurity self-assessment tool developed in the Kytee project to strengthen its cybersecurity capabilities and gain a clearer understanding of areas for improvement. Although information security was already an established part of the company's operations, the pilot highlighted new opportunities for further development.
The greatest benefit of the pilot was identifying concrete next steps. The tool helped structure the broad field of cybersecurity, improved understanding of the connections between different security domains, and provided a solid foundation for continuous improvement. It also encouraged regular discussions about cybersecurity as part of the company's day-to-day operations.
According to Knobbi Oy CEO Tommi Dahlroos, the self-assessment tool provided a ready-made framework that made evaluating the company's cybersecurity maturity much easier than starting from scratch. The integrated AI chat feature also helped users interpret the questions and made the assessment process more straightforward.
The pilot reinforced the idea that cybersecurity is not a one-time project but an ongoing process that requires regular attention. The tool encouraged participation across the organization, making cybersecurity a shared responsibility rather than the task of a single individual.
According to Kytee project manager Jere Grönman, the self-assessment tool has since been independently audited by a third party, further strengthening its credibility. During the project, the decision was made to make the tool freely available to all companies, enabling organizations to improve their cybersecurity practices and prepare for future regulatory requirements.
Knobbi Oy's experience demonstrates that a well-designed self-assessment tool can make cybersecurity more accessible and practical while helping organizations identify development priorities and integrate cybersecurity into their everyday business.
RoboAI Academy
RoboAI Academy students developed a VR-based safety training demo for Pori Energia to make workplace safety training more engaging, interactive, and effective. The project focused on creating an immersive learning experience that supports practical safety training.
The virtual environment allows employees to practice safety procedures, identify workplace hazards, and respond to emergency situations in a realistic but completely safe setting. By combining 360° imagery, virtual reality, and gamified learning elements, the solution encourages active participation and improves learning outcomes.
The project gave students valuable hands-on experience in VR development, scenario-based learning, and user experience design while delivering a functional proof-of-concept for the company.
Favorit Tuote Oy was looking for an affordable and easily replicable solution to ensure that the correct components are included in every accessory package. RoboAI Academy students developed two prototype systems to reduce packaging errors and support assembly work.
The project resulted in two Arduino-controlled prototypes that use ultrasonic sensors and LED lights to guide operators through the packing process. The two alternative designs enabled the company to evaluate different implementation approaches for future production use.
The project provided students with practical experience in automation, sensor technology, programming, and product development while delivering two functional proof-of-concept solutions for the company.
RoboAI Academy students explored how the palletizing of large cardboard boxes at Tukimet Oy could be automated using a collaborative robot. The project aimed to reduce physical strain, improve ergonomics, and evaluate the suitability of robotics for the company's production environment.
The production area was laser-scanned, and a detailed 3D simulation model was created using Visual Components software. Several collaborative robot concepts were simulated and compared in terms of reach, safety, cycle times, and overall feasibility within the existing production layout.
As a result, Tukimet Oy received practical recommendations to support future automation planning, while the students gained valuable hands-on experience in laser scanning, robotics, and industrial 3D simulation.
RoboAI Academy students explored how the physically demanding task of portioning food from cooking kettles into transport containers could be improved through robotics and automation for the City of Pori central kitchen. The project focused on enhancing workplace ergonomics while evaluating the feasibility of robotic solutions in a real commercial kitchen.
The kitchen was laser-scanned to create a detailed 3D simulation model, which was used to design and evaluate three different robotic concepts. The simulations assessed each solution's space requirements, safety, performance, and usability within the food portioning process.
The results showed that full automation of the task would not be the most practical solution, but the project identified several realistic ways to improve ergonomics. Students gained valuable hands-on experience in robotics and 3D simulation, while the customer received practical recommendations to support future development.
RoboAI community cases
The pilot site for the Robocoast project was Cimcorp Oy in Ulvila. In the first phase of the pilot, data related to Cimcorp's Dream Factory concept was studied.
Data from the actuators is measured throughout the operation to ensure quality and to support product development. Data analytics will be used to analyze the processes of the actuators and optimize some aspects to (further) improve performance. In analytics, data generated by the devices is retrieved from a database and analyzed to infer the performance of both individual devices and the overall system, and to make the necessary changes (iteratively) to the control system.
In addition to iteration, there are numerous opportunities to optimize production by optimizing, for example, incorrect motion paths from robots, such as too much time/current, the arc/angle of a moving actuator being too large, and the use of histograms to find fault conditions.
Data analytics can also be used to create parallel metrics by checking the reliability of the data using data from multiple data sources. This allows early detection of possible failure or malfunction of an actuator/component.
The second phase of the pilot focused on analyzing the data from the box inventory management system. The development work experimented with different data visualization tools, mainly for maintenance and development purposes. At the same time, however, it had to be kept in mind that the same data could be added in the future for customer service purposes in a clear and coherent way.
The pilot ran for a year and ended in December 2020. During the pilot, it became clear that data analysis work is a long-term and ever-changing process in an ever-evolving operator environment. As new features and new metrics are added to actuators, data analytics must keep pace and provide customers with up-to-date, clearly structured information.
Cimcorp Oy produces automated actuators for the industry for inventory management and material handling. Learn more about Cimcorp. (cimcorp.com)
The City of Pori pilot aimed to support the design and implementation of the remote work experience component of the biennial well-being at work survey by analysing and visualising the data. An expert from the DAO research group at the Tampere University developed the survey platform, which in practice consisted of two parts: semi-structured interviews and an electronic survey. The topic was very topical in view of the current situation and the aim of the survey was to get a fresh perspective from the expert of the research team.
It was important to find out how remote work is perceived in terms of well-being at work, what challenges it may have faced, and how it has affected the overall well-being of the employee. In particular, remote working remained a more permanent way of working even after the epidemic.
The semi-structured interviews involved 20 participants, of whom 6 were in a pre-personal position; 12 women and 8 men. With an average age of around 45 years; interview duration of around 25-45min; interviewees represented a wide range of industries and different types of jobs. The interview material was transcribed and analysed qualitatively using thematic data analysis.
The written report will describe in more detail the differences between sectors, but the data analysis already shows that they follow the same pattern as the results of the online survey.
A total of 922 employees of the City of Pori responded to the teleworking survey, 69% of whom had not teleworked before. Before the teleworking recommendation, the average number of days teleworking was 0.22 per week. The increase in teleworking averaged 2.27 days per week.
The Tampere University Pori unit also does a lot of business cooperation in the form of commissioned case studies. For Honkajoki Oy, a reliable automated analysis mechanism for slaughterhouse waste based on hyperspectral imaging is being developed, which could be used for further processing of slaughterhouse waste.
The unit was equipped with a hyperspectral camera from Specim Oy, which allows the detection of the level of contamination and contaminants such as metal, plastic, glass and water content in the slaughterhouse waste from the line. The hyperspectral data reference values are taken from laboratory samples of slaughterhouse waste, which have been jointly imaged by researchers from the Tampere University and Specim Oy.
The hyperspectral camera measurements are fed into the Tampere University’s neural network (artificial intelligence) computing unit, where the neural network tries to learn a cause-and-effect relationship between the hyperspectral camera measurements and the laboratory measurements of slaughterhouse waste. Several intermediate steps are needed to teach the neural network, as there is a lot of hyperspectral data and before teaching the neural network, the amount of data needs to be reduced without losing any essential data. Then, the size of the neural network needs to be determined and the optimal network learning parameters need to be found before the actual neural network teaching.
In addition, Dyme Solutions Oy has implemented a data collection solution to collect data from different processes at the Honkajoki Oy production plant in one place. The researchers at the Tampere University aim to find cause and effect relationships in the collected data, which can be used to take development measures for different processes, for example to look at energy consumption and in the future to reduce energy consumption or to predict the degradation of process-related components.
