Jim Tørresen

• Nishimura, Markus Yuki; Matsumoto, Kohei; Tørresen, Jim & Kurazume, Ryo (2026). FLexible Multi-Robot Social Navigation Using Diffusion-Based Deep Reinforcement Learning. In Ishii, Kazuo; Shibata, Tomohiro; Howard, Ayanna; Rossi, Silvia & Clodic, Aurélie (Ed.), Proceedings of The 35th IEEE International Conference on Robot and Human Interactive Communication (RO-MAN 2026). IEEE. ISSN 9798331588656. p. 2122–2127. doi: 10.1109/RO-MAN72286.2026.11714669. Full text in Research Archive Show summary

  Diffusion-based reinforcement learning has recently shown strong performance in robot navigation, and prior work has demonstrated that Q-Score Matching (QSM) is effective in crowded environments. Building on this line of work, we study inference-time guidance for multi-robot social navigation using Multi-Agent Q-Score Matching (MAQSM) as the backbone policy. Rather than modifying the training objective, the proposed method steers reverse diffusion sampling of a trained MAQSM policy at evaluation time toward more coordinated multi-robot behavior. The guidance combines formation consistency and goal alignment, enabling deploymenttime control without retraining. In our implementation, robots share the same diffusion policy parameters, generate per-robot actions under this shared policy, and execute the resulting multi-robot action jointly in the simulator. Experiments first compare QSM and MAQSM to motivate the choice of MAQSM as the underlying multi-robot backbone, and then compare unguided and guided MAQSM in a more challenging bidirectional scenario. The results show that inference-time guidance improves success and safety in denser settings, at the cost of longer navigation time. These findings position post-training guidance as a practical way to improve diffusion-based multirobot policies while preserving the original learning pipeline.

• Tørresen, Jim; Lindblom, Diana Saplacan; Baselizadeh, Adel; Mahler, Tobias & Bygrave, Lee Andrew (2026). Reducing Robot Vulnerabilities Through Joint Regulatory Assessment and User-Centered Development and Testing. In Weng, Yueh-Hsuan (Eds.), Social Robots and Ethical Design. Springer. ISSN 9789819225156. Full text in Research Archive Show summary

  This book chapter introduces research examples from a unique collaboration between legal scholars and technology scholars who have together endeavored to make progress in safe and privacy-preserving robot assistant development and deployment. This collaboration has taken place within the “Vulnerability in the Robot Society” (VIROS) research project at the University of Oslo, Norway (VIROS 2025). Institutionally, this is a joint research project between the Norwegian Research Center for Computers and Law (NRCCL) at the Department of Private Law and the Robotics and Intelligent Systems (ROBIN) research group at the Department of Informatics. The project explores challenges regarding digital vulnerabilities that the adoption of robots in human environments poses, particularly as regards privacy, safety, and security, and how these challenges can be met from legal and technical perspectives. We have been particularly interested in exploring assistant and healthcare robots since this domain imposes high demands on the technology to be used. As part of the project, we have analyzed various European Union (EU) regulations that have been introduced in recent years, with a view to assessing their impact in the domain. The TIAGo robot from PAL Robotics has been the main robot platform for technical work and experimental studies in the project. A number of novel methods for robot sensing and control have been implemented. Further, several of them have been assessed through human–robot interaction studies. This chapter introduces the VIROS project and outlines some of our regulatory and technological research. We also highlight several of the conflicting trade-offs seen in the domain of HRI (Torresen et al. 2023).

• Lindblom, Diana Saplacan; Kuoppamäki, Sanna; Bodenhagen, Leon & Tørresen, Jim (2026). Robots as Welfare Technologies and Actors within Future Homeand Healthcare Services. In Havinen, Laura & Hägglund, Susanne (Ed.), Adjunct Proceedings of the 14th Nordic Conference on Human-Computer Interaction. Association for Computing Machinery (ACM). ISSN 9798400727986. Full text in Research Archive Show summary

  This workshop explores whether, how, and under which conditions Welfare Robots (WRs), may be meaningfully integrated into future home- and healthcare services and practices. Building on current debates on welfare technology, AI-driven robots, and post-growth Human-Robot Interaction (HRI), the workshop invites designers, roboticists, engineers, healthcare professionals, philosophers, sociologists, and other practitioners to critically examine and systematically engage in contributing to exploring design alternatives and co-design a future model for responsible integration of robots into future care services, using Story Dialogue Method.

• Lindblom, Diana Saplacan; Weng, Yueh-Hsuan; Bravo, Pedro Pablo Lucas; Baselizadeh, Adel & Tørresen, Jim (2026). A Cross-Cultural Video-Based Study between Norway and Japan on Informed Consent and Disclosure Mechanisms on the Use of Social Robots in Public and Private Spaces - Users' Perspectives. In Ishii, Kazuo; Shibata, Tomohiro; Howard, Ayanna; Rossi, Silvia & Clodic, Aurélie (Ed.), Proceedings of The 35th IEEE International Conference on Robot and Human Interactive Communication (RO-MAN 2026). IEEE. ISSN 9798331588656. Full text in Research Archive
• Sikora, Claudia Magdalena; Otterdijk, Marieke van; Soma-Jestilä, Rebekka; Tørresen, Jim & Lindblom, Diana Saplacan (2026). "There’s something human about it!” - Exploring users’ perception of nonverbal communication of robotic furniture. In Samani, Hooman & Akhond, Saina (Ed.), Proceedings of the International Confernece on Social Robotics (ICSR) + ART 2026. Springer International Publishing . Full text in Research Archive Show summary

  Non-anthropomorphic robots are integrating into our homes. People attribute human-like characteristics to these robots, although they do not resemble humans. This raises the question of whether similar processes would occur when people interact with robot furniture. Thus, the presented exploratory pilot study investigates how users intuitively interpret and perceive the non verbal communication of such robots. In an instrumental case study, we had eleven participants interacting with the T-ABLE robot through expressive behavior. Participants were interviewed about their experience. Our findings indicate that the participants use their associations between behavior and prior to similar experiences to attribute personality to the robot table. These findings provide insights for theorizing the intuitive comprehension of the behavior of non-anthropomorphic robots.

• Sarfraz, Burhan Mohammad; Lindblom, Diana Saplacan; Baselizadeh, Adel & Torresen, Jim (2026). The Privacy-Preserving Capabilities of a Service Robot in a Scenario-Based Healthcare Setting. In Baillie, Lynne; Smart, William D.; Graaf, Maartje De; Gombolay, Matthew & Torre, Ilaria (Ed.), Proceedings of the 21st ACM/IEEE International Conference on Human-Robot Interaction (HRI Companion '26). Association for Computing Machinery (ACM). ISSN 9798400723216. p. 675–679. doi: 10.1145/3776734.3794481. Full text in Research Archive
• Lømo, Tobias; Baselizadeh, Adel; Ellefsen, Kai Olav & Tørresen, Jim (2026). Dual Process Dreamer: Fast and Slow Decision-Making with World Models. Proceedings of the International Conference on Agents and Artificial Intelligence (ICAART). ISSN 2184-3589. 2, p. 1230–1241. doi: 10.5220/0014243200004052. Full text in Research Archive Show summary

  Most robot systems are based on a single decision-making process. This process needs to balance time, energy, and accuracy in every situation. However, according to ”dual process theory” (DPT) from cognitive psychology, this is not how humans work. Depending on the situation, we have the ability to switch between two thinking methods, a fast system 1 (S1) and a slower system 2 (S2). In this paper, we propose a novel approach to a dual process architecture for robots and agents. Our method, called Dual Process Dreamer (DPDreamer), is a combination of a reinforcement learning policy network, a planning algorithm, and a learned world model. The world model allows the parts of DPDreamer to work together and create a more integrated system compared to previous proposals of DPT systems. DPDreamer was tested in a puzzle game called Sokoban, and by balancing the use of S1 and S2, DPDreamer managed a success rate similar to S2 while using S1 most of the time, showing the benefit of using a more adaptable system.

• Guo, Jinyue; Tørresen, Jim & Jensenius, Alexander Refsum (2026). Investigating Auditory–Visual Perception Using Multi-Modal Neural Networks with the SoundActions Dataset. Transactions of the International Society for Music Information Retrieval. 9(1), p. 85–85. doi: 10.5334/tismir.223. Full text in Research Archive Show summary

  Musicologists, psychologists, and computer scientists study relationships between auditory and visual stimuli from very different perspectives and using various terminologies and methodologies. This article aims to bridge the gap between phenomenological sound theory, auditory–visual theory, and audio–video processing and machine learning. We introduce the SoundActions dataset, a collection of 365 audio–video recordings of (primarily) short sound actions. Each recording has been human‑labeled and annotated according to Pierre Schaeffer’s theory of reduced listening, which describes the property of the sound itself (e.g., ‘an impulsive sound’) instead of the source (e.g., ‘a bird sound’). With these reduced‑type labels in the audio–video dataset, we conducted two experiments: (1) fine‑tuning the latest audio–video transformer model on the reduced‑type labels in the SoundActions dataset, proving that the model can recognize reduced‑type labels, and observing that the modality‑imbalance phenomenon is similar to the added value theory by Michel Chion and (2) proposing the Ensemble of Perception Mode Adapters method inspired by Pierre Schaeffer’s three listening modes, improving the audio–video model also on reduced‑type tasks.

• Iwashita, Yumi; Moe, Haakon; Cheng, Yang; Ansar, Adnan; Georgakis, Georgios & Stoica, Adrian [Show all 9 contributors for this article] (2025). Enhancing the Quality of 3D Lunar Maps Using JAXA’s Kaguya Imagery. In Nasreddine, Belkacem Abdelkader (Eds.), 2025 IEEE International Conference on Systems, Man, and Cybernetics (SMC). IEEE (Institute of Electrical and Electronics Engineers). ISSN 9798331533588. p. 763–768. doi: 10.1109/SMC58881.2025.11343204. Full text in Research Archive
• Paulsen, Geir; Cardenas, Juan Sebastian; Alsgaard, Rosa Nicoline Pham; Baselizadeh, Adel; Uddin, Md Zia & Tørresen, Jim (2025). AI-Based User Gesture Recognition for Human-Robot Interaction Using Wrist Sensors. In Kita, Eisuke (Eds.), 2025 11th International Conference on Robotics and Artificial Intelligence (ICRAI), December 19-21, 2025, Nagoya, Japan. IEEE (Institute of Electrical and Electronics Engineers). ISSN 9798331590680. p. 161–165. doi: 10.1109/icrai68431.2025.11396715. Full text in Research Archive Show summary

  Natural and intuitive interaction remains a central challenge in human-robot communication. This study presents a privacy-preserving gesture recognition framework that enables command-based interaction through simple hand and arm movements. Motion data are captured using a wrist-worn sensor, eliminating the need for cameras or other intrusive tracking systems. Data from ten participants performing seven gesture classes, including move forwards, move backward, move left-right, spin horizontally, spin vertically, wave, and move up-down are used to train and validate a hybrid Convolutional Long Short-Term Memory (CNN-LSTM) model. An additional non-command state is included to ensure the robot remains inactive when no gesture is detected. The proposed model achieves an overall recognition accuracy of approximately 95% across all gesture classes using cross-validation. This framework enhances the safety, intuitiveness, and fluidity of human-robot interaction and provides a robust foundation for gesture-based robot control and human activity recognition.

• Tønnessen, Håkon Haustreis; Uddin, Md Zia & Tørresen, Jim (2025). PINE: Planning and Identifying Neural Network for Thinking Fast and Slow. In Kita, Eisuke (Eds.), 2025 11th International Conference on Robotics and Artificial Intelligence (ICRAI), December 19-21, 2025, Nagoya, Japan. IEEE (Institute of Electrical and Electronics Engineers). ISSN 9798331590680. p. 191–197. doi: 10.1109/icrai68431.2025.11396706. Full text in Research Archive Show summary

  The human brain's unique ability to seamlessly switch between quick, effortless reactions and slow, deliberate thinking forms the foundation for the content of this paper. Humans effortlessly navigate complex tasks like walking, speaking, and driving, relying on heuristic responses to their environment. This natural division of labor allows humans to adapt quickly and efficiently, tackling both new challenges and familiar tasks. Drawing inspiration from this cognitive duality, we propose, in this paper, a machine-based analog that employs a two-level functionality mirroring human capabilities. The approach centers on a Reinforcement Learning agent that replicates human learning through trial and error. The thesis introduces the Planning and Identifying Neural Network (PINE), a framework that identifies tasks using a combination of inherited environmental information, observations, and human-readable task descriptors. PINE leverages a model-based Reinforcement Learning algorithm, streamlining agent complexity and significantly reducing solution-finding time compared to model-free algorithms. PINE operates in real-time, utilizing initial environmental observations and task descriptors to identify and categorize encountered tasks. For unknown tasks, PINE employs meta-learning, selecting the most similar known task to expedite the solution-finding process with fewer episodes. Despite the challenge of task identification without interaction or labels, PINE demonstrates high accuracy in recognizing previously solved tasks and providing reliable estimates for the most similar tasks. This innovative two-level learning model represents a significant step towards creating adaptive machines that efficiently handle both familiar and novel challenges.

• Baselizadeh, Adel; Uddin, Md Zia; Khaksar, Weria; Lindblom, Diana Saplacan & Tørresen, Jim (2025). Privacy-Preserving 3D Lidar-Based Multi-Modal Activity Recognition in Human-Robot Interaction. International Conference on Control, Mechatronics and Automation. ISSN 2837-5114. 2025, p. 516–523. doi: 10.1109/iccma67641.2025.11369549. Full text in Research Archive Show summary

  Human activity recognition (HAR) involves using sensors to collect human data, which is then analyzed to identify their activities. HAR has numerous applications across various fields. One significant area is human-robot interaction (HRI). Recognizing user activities provides robots with valuable insights into the user’s status, enhancing the efficiency of HRI. However, developing HAR models presents privacy challenges due to the collection of user data, especially in robots, which are often equipped with diverse sensors. This paper proposes privacy-preserving HAR methods within the context of HRI.The paper investigates using privacy-preserving sensors, with a particular focus on 3D Lidar, to develop HAR models. It explores the integration of 3D Lidar with other sensors, including user-wearable sensors and robot-based sensors, including force/torque sensors, through multimodal deep learning (DL) approaches. Various DL-based sensor fusion methods, including data-level and feature-level fusion approaches, are thoroughly examined, and their accuracies for HAR are compared.A novel dataset was collected to train the multimodal DL models, capturing various user activities during HRI. This dataset leverages 10 different sensors, including 9 privacy-preserving sensors, along with an RGB camera for reference. The paper considers nine distinct user activities, including physical interactions with a robot and commanding a robot to perform specific tasks. The results indicate that integrating the sensors’ data at the feature-level achieves an 80.73% accuracy in recognizing various user activities during HRI.

• Santos, Letícia dos; Tørresen, Jim; Kolberg, Mariana & Maffei, Renan (2025). An Autonomous Floor Clearing Strategy to Tidy up Unknown Home Environments with a Mobile Manipulator Robot. In Slawiñski, Emanuel; García, Cecilia & Tosetti, Santiago (Ed.), 2025 IEEE International Conference on Advanced Robotics (ICAR). IEEE (Institute of Electrical and Electronics Engineers). ISSN 9798331578107. doi: 10.1109/icar65334.2025.11338715. Full text in Research Archive
• Torresen, Jim; Laschi, Cecilia; Prestes, Edson; Kavraki, Lydia E.; Caleb-Solly, Praminda & Weng, Yueh-Hsuan (2025). Robot Ethics: Ethical, Legal, and User Perspectives in the Development and Application of Robotics and Automation [From the Guest Editors]. IEEE robotics & automation magazine. ISSN 1070-9932. 32(4), p. 11–11. doi: 10.1109/mra.2025.3619589. Full text in Research Archive
• Pham, Hoang Minh; Noori, Farzan Majeed; Uddin, Md Zia & Tørresen, Jim (2025). Heart Rate Forecasting Using Ultra-Wideband Radar with Sequence-to-Sequence Model. In Cruz, Luis; Loncaric, Sven; Chen, Zhibo; Pinho, Armando J.; Batista, Jorge & Subasic, Marko (Ed.), Proceedings of the 14th International Symposium on Image and Signal Processing and Analysis (ISPA), Coimbra, Portugal, October 29-31, 2025. IEEE (Institute of Electrical and Electronics Engineers). ISSN 9798331577551. p. 16–21. doi: 10.1109/ispa66905.2025.11259452. Full text in Research Archive Show summary

  Heart dysfunction can be very dangerous and directly associated with the risk of fatality. Therefore, automatic heart rate monitoring is important for detecting a heart abnormality early and providing timely intervention. The development of advanced sensor technology allows continuous tracking of the heart rate, as well as providing more health-related data sources, which could further improve the forecasting performance. In this study, we investigate the use of recent heart rate values and Ultra-Wideband (UWB) data in the last 20s to predict the next 10s of heart rate. Data were collected from 20 participants performing different activities. We examine the forecasting by using only a heart rate sensor (uni-modality) and by using heart rate sensor and UWB together (multi-modality). Three levels of activity were included in the study: resting (rest), lying after doing exercises (lying-abnormal) and lying after full recovery (lying-normal). A multi-modality Sequence-to-Sequence (Seq2Seq) model has a 10-percent quantile losses of 5.20 for rest activity, 4.92 for lying-abnormal activity, and 4.24 for lying-normal activity. With the lowest loss, Seq2Seq outperforms Autoregressive Integrated Moving Average (ARIMA), Error Trend and Seasonality (ETS), Transformer and Temporal Fusion Transformer (TFT) in forecasting heart rate in the three activity levels. Our results point to a new direction in multimodal prediction of heart rate to assist healthcare, using deep learning approaches.

• Goka, Shunsuke; Strand, Ørjan; Miura, Jun & Tørresen, Jim (2025). Situation-Based Navigation Strategy Switching for Mobile Robots in Dynamic Pedestrian Environments. In Uchiya, Takahiro; Nishimura, Ryota & Ikehara, Tadaaki (Ed.), 2025 IEEE 14th Global Conference on Consumer Electronics (GCCE). IEEE (Institute of Electrical and Electronics Engineers). ISSN 9798331524166. doi: 10.1109/gcce65946.2025.11275115. Full text in Research Archive Show summary

  This paper proposes a method for adaptively switching navigation strategies based on a visual assessment of the current situation. Despite recent progress in motion planning and navigation methods, it is still challenging to adopt a single navigation method in various situations, which are determined by both the environment structure and pedestrian distribution and movements. A promising approach is to switch between multiple navigation strategies depending on the current situation. The proposed method adopts a visual situation classification, trained on an automatically labeled dataset, for choosing an appropriate local motion planner. Experimental comparison with fixed planners using a real robot demonstrates the superiority of our method in terms of both navigation efficiency and safety.

• Watanabe, Shin; Horn, Geir; Tørresen, Jim & Ellefsen, Kai Olav (2025). Integrating Bilevel Planning and Offline Skill Learning for Enhancing Mobile Manipulation, 2025 IEEE 21st International Conference on Automation Science and Engineering (CASE). Institute of Electrical and Electronics Engineers (IEEE). ISSN 9798331522476. p. 2275–2280. doi: 10.1109/CASE58245.2025.11163881. Full text in Research Archive Show summary

  Solving complex robotic mobile manipulation tasks requires both planning the sequence of skills to execute and learning how to robustly execute each skill. Planning-based approaches such as task and motion planning (TAMP) can help train skills more efficiently through demonstrations, while learning-based approaches such as reinforcement learning (RL) can help plan tasks more quickly through heuristics. This paper presents a novel approach to generalizing mobile manipulation tasks by synergistically combining sampling-based TAMP and value-based RL. The TAMP solver first generates suboptimal demonstration trajectories of a particular skill, from which an offline RL algorithm distills a robust policy and a value function, the latter serving as a skill feasibility classifier. The policy and the classifier are both fed back into the TAMP workflow not only to improve the skill success rate but also to speed up the planner by sampling robot configurations for which the skill is likely to succeed. We evaluate the approach on a simulated block-pushing domain. Re-purposing a byproduct of an offline skill-learning process leads to an integrated planning and learning system that exploits the awareness of its own skill competence.

• Moghaddam, Emil Engelstad; Uddin, Md Zia; Miura, Jun & Tørresen, Jim (2025). A Robust Approach for Motion Skill-Based Scene Categorization. In Neri, Filippo (Eds.), Proceedings of the 2025 10th International Conference on Machine Learning Technologies (ICMLT), May 23-25, 2025, Helsinki, Finland. IEEE (Institute of Electrical and Electronics Engineers). ISSN 9798331536725. p. 482–487. doi: 10.1109/icmlt65785.2025.11193163. Full text in Research Archive Show summary

  This paper introduces a novel method for autonomous navigation of mobile robots, combining functional scene categorization with tailored navigation strategies. Scene categorization is a computer vision task that identifies environmental types in images. Our approach establishes a taxonomy for scene categorization with the aim of utilization in mobile robots. We used a motion-skill dataset for training different neural networks. The dataset, comprising 19923 images with nine navigation strategy labels, demonstrated the feasibility of coupling environment recognition with neural networks. Three architectures (ResNet-50, BEiT-Base, ConvNeXt-Tiny) achieved notable performance, with ConvNeXt-Tiny leading with a weighted F1 score of 0.934. Pre-training on the SUN397 dataset significantly improved the model performance, emphasizing the importance of leveraging existing scene categorization datasets. This study highlights the synergy between scene categorization and functional scene categorization, providing insights for designing efficient, context-aware autonomous navigation systems.

• Guo, Jinyue; Tørresen, Jim & Jensenius, Alexander Refsum (2025). Cross-modal Analysis of Spatial-Temporal Auditory Stimuli and Human Micromotion when Standing Still in Indoor Environments. In McArthur, Angela; Matthews, Emma-Kate & Holberton, Tom (Ed.), Proceedings of the 17th International Symposium on Computer Music Multidisciplinary Research. The Laboratory PRISM “Perception, Representations, Image, Sound, Music”. ISSN 9791097498061. p. 871–882. doi: 10.5281/ZENODO.17502603. Full text in Research Archive Show summary

  This paper examines how a soundscape influences human stillness. We are particularly interested in how spatial and temporal features of a soundscape influence human micromotion and swaying patterns. The analysis is based on 345 Ambisonics audio recordings of different indoor environments and corresponding accelerometer data captured at the chest of a person standing still for ten minutes. We calculated the temporal and spatial correlation between the person's quantity of motion and the sound energy of the Ambisonic recordings. While no clear temporal correlations were found, we discovered a correlation between the spatial directionality of the micromotion and the sound direction of arrival. The results suggest a potential entrainment between the directionality of environmental sounds and human swaying patterns, which have not been thoroughly studied previously compared to the temporal or spectral features of indoor soundscapes.

• Nergård, Katrine Linnea; Ellefsen, Kai Olav & Tørresen, Jim (2025). Fast or Slow: Adaptive Decision Making in Reinforcement Learning with Pre-Trained LLMs. In Ugur, Emre; Sciutti, Alessandra & Rohlfing, Katharina (Ed.), 2025 IEEE International Conference on Development and Learning (ICDL). IEEE (Institute of Electrical and Electronics Engineers). ISSN 9798331543433. doi: 10.1109/ICDL63968.2025.11204357. Full text in Research Archive
• Meijer, Frida; Lindblom, Diana Saplacan; Baselizadeh, Adel & Tørresen, Jim (2025). "The wooden gripper was warmer and made the robot less threatening"– A Study on Perceived Safety based on Robot Gripper’s Visual and Tactile Properties. In Emilia, Barakova,; Ben, Allouch, Somaya; Kazuhiro, Nakadai, & Goldie, Nejat, (Ed.), Proceedings of the IEEE International Conference on Robot and Human Interactive Communication (RO-MAN) 2025. IEEE (Institute of Electrical and Electronics Engineers). ISSN 9798331587710. p. 1091–1098. doi: https:/ieeexplore.ieee.org/document/11217775. Full text in Research Archive Show summary

  An ageing population and the need of providing adequate care have led to developing robots to relieve healthcare workers and to assist individuals in their own homes. However, the successful integration of robots in such settings relies on more than just ensuring physical safety associated with physical risks (e.g., collisions): it also requires the user’s perceived safety – the users perceiving the robot as not doing any harm. This paper explores the potential influence of a robot grippers’ visual and tactile properties, such as materials and texture, on the users’ perceived safety and comfort of human-robot interaction. An initial survey was distributed to 53 participants, exploring five (n=5) robot gripper designs focusing on the robots’ gripper shape. One design shape was thereafter selected to be constructed as a cover to be placed over the parallel grippers of the TIAGo robot, by using 1) wood filament and 2) plastic. The covers were then tested in an experimental setting with 11 participants. The covers were attached to the TIAGo mobile manipulator robot and participants interacted with both of the designed gripper covers within a controlled laboratory environment. A questionnaire was distributed to all 11 experiment participants, at different stages of the interactions. The findings indicate that the material of the gripper influenced participants’ sense of comfort, familiarity, and perceived capabilities of the robot. The study suggests that perceived safety in human-robot interaction (HRI) is shaped not only by physical factors but also by how materials are personally and contextually interpreted. To better support safe and comfortable interactions, further research is needed to understand how material choices shape users’ perceived safety.

• Otterdijk, Marieke van; Neggers, Margot; Torresen, Jim & Barakova, Emilia (2025). Exploring Human Attribution of Emotional Intent to Motion Features in a Humanoid Robot. In Palinko, Oskar; Bodenhagen, Leon; Cabibihan, John-John; Fischer, Kerstin; Šabanović, Selma; Winkle, Katie; Behera, Laxmidhar; Ge, Shuzhi Sam; Chrysostomou, Dimitrios; Jiang, Wanyue & He, Hongsheng (Ed.), Social Robotics. Springer. ISSN 9789819635184. p. 323–323. doi: 10.1007/978-981-96-3519-1_29. Full text in Research Archive
• Maeda, Ryuichi; Baselizadeh, Adel; Watanabe, Shin; Kurazume, Ryo & Tørresen, Jim (2025). Adaptive Tidying Robots: Learning from Interaction and Observation. In Asfour, Tamim; Ramírez-Amaro, Karinne; Kim, Joohyung & Cheng, Gordon (Ed.), 2025 IEEE/SICE International Symposium on System Integration (SII). IEEE (Institute of Electrical and Electronics Engineers). ISSN 9798331531614. p. 185–192. doi: 10.1109/sii59315.2025.10871003. Full text in Research Archive
• Otterdijk, Marieke van; Laeng, Bruno; Lindblom, Diana Saplacan; Baselizadeh, Adel & Tørresen, Jim (2025). Seeing Meaning: How Congruent Robot Speech and Gestures Impact Human Intuitive Understanding of Robot Intentions. International Journal of Social Robotics. ISSN 1875-4791. 17, p. 2279–2292. doi: 10.1007/s12369-025-01271-0. Full text in Research Archive Show summary

  Social communication between humans and robots has become critical as a result of the integration of robots into our daily lives as assistants. There is a need to explore how users intuitively understand the behavior of a robot and the impact of social context on that understanding. This study measures mental effort (as indexed by pupil response) and processing time, measured as the time taken to provide the correct answer, to investigate participants’ intuitive understanding of the robot’s gestures. Thirty-two participants participated in a charades game with a TIAGo robot, during which their eyes were tracked. Our findings show a relationship between mental effort and processing time, and indicate that robot gestures, congruence of speech and behavior, and the correctness of interpreting robot behavior influence intuitive understanding. Furthermore, we found that people focused on the robot’s limb movement. Using these findings, we can highlight what features contribute to the intuitive interaction with a robot, thus improving its efficiency.

• Weng, Yueh-Hsuan; Torabi, David; Tørresen, Jim; Dong, Zonghao & Hirata, Yasuhisa (2025). Bridging Ethics and Reality: Integrating Thought Experiments and Empirical Insights in Robot Ethics. IEEE robotics & automation magazine. ISSN 1070-9932. 32(4), p. 84–90. doi: 10.1109/mra.2025.3584352. Full text in Research Archive
• Orten, Kristine Fjellkårstad; Helgesen, Sander Elias Magnussen; Chen, Bihui; Baselizadeh, Adel; Tørresen, Jim & Herrebrøden, Henrik (2025). Can machine learning distinguish between elite and non-elite rowers? International Journal of Computer Science in Sport. 24(1), p. 118–132. doi: 10.2478/ijcss-2025-0007. Full text in Research Archive Show summary

  A major challenge for sports coaches and analysts is to identify critical elements of athletes’ movement patterns. A potentially relevant tool is machine learning, useful because of its ability to extract patterns from data. In the current study, we employed various deep learning frameworks, including Gated Recurrent Unit networks (GRUs), Convolutional Neural Networks (CNNs), and Multi-Layer Perceptrons (MLPs), to search for differences between elite and non-elite rowers using a rowing ergometer. The MLP model achieved an accuracy of 100% when using all input features, indicating that the problem is suitable as a machine learning task. Our research focused on using a limited amount of the data. Despite using fewer input features, the models managed to classify skill levels with reasonable precision, reaching a best performance of 77% accuracy for the model combining GRU and CNN architectures, 78% for the GRU model, and 94% for the MLP model. From a rowing perspective, the results suggest that movement coordination between upper and lower body limbs, as represented by different feature combinations, is informative in distinguishing between elites and non-elites. The current work suggests that machine learning may supplement human experts in sports coaching, analytics, and talent identification.

• Duric, Adrian; Tørresen, Jim; Riegler, Michael & Hammer, Hugo Lewi (2025). Explanation Supported Learning: Improving Prediction Performance with Explainable Artificial Intelligence. IEEE International Symposium on Computer-Based Medical Systems. p. 591–598. doi: 10.1109/CBMS65348.2025.00125. Full text in Research Archive Show summary

  When artificial intelligence (AI) and machine learning (ML) models are applied in healthcare, the ability to understand and explain model decisions is an important aspect. Methods in the field of explainable AI (XAI) have been developed to create explanations for such decisions, which provides transparency and trust to the prediction model. However, the use of XAI-based explanations as added data features for the purpose of improving prediction performance remains a little explored topic. Our proposed Explanation Supported Learning (XSL) framework can improve classification performance for ML models used in medical imaging systems, while also providing a new understanding of how medical images are processed by deep learning (DL) models. The XSL framework consists of novel methods to achieve knowledge transfer from one or several teacher models to a student model. The novelty lies in using explanations from the teacher models, obtained from XAI techniques, as added features when training the student model. This approach enables flexible knowledge transfer between models of different architecture types. We further demonstrate how the XSL framework can be used as a new metric for measuring the quality of the explanations provided by XAI methods. The achievement of increased performance in this framework requires that the chosen XAI technique contains useful information based on the learned understanding of the input data by the teacher models. By testing XSL on the HyperKvasir gastrointestinal image dataset, we achieved significant increases in most of the measured classification metrics, and exceeded most benchmark scores of the HyperKvasir paper. Our code is available on GitHub.

• Fuhrer, Julian; Glette, Kyrre; Ivanovic, Jugoslav; Larsson, Pål Gunnar; Bekinschtein, Tristan & Kochen, Silvia [Show all 11 contributors for this article] (2025). Direct brain recordings reveal implicit encoding of structure in random auditory streams. Scientific Reports. 15. doi: 10.1038/s41598-025-98865-5. Full text in Research Archive Show summary

  The brain excels at processing sensory input, even in rich or chaotic environments. Mounting evidence attributes this to sophisticated internal models of the environment that draw on statistical structures in the unfolding sensory input. Understanding how and where such modeling proceeds is a core question in statistical learning and predictive processing. In this context, we address the role of transitional probabilities as an implicit structure supporting the encoding of the temporal structure of a random auditory stream. Leveraging information-theoretical principles and the high spatiotemporal resolution of intracranial electroencephalography, we analyzed the trial-by-trial high-frequency activity representation of transitional probabilities. This unique approach enabled us to demonstrate how the brain automatically and continuously encodes structure in random stimuli and revealed the involvement of a network outside of the auditory system, including hippocampal, frontal, and temporal regions. Our work provides a comprehensive picture of the neural correlates of automatic encoding of implicit structure that can be the crucial substrate for the swift detection of patterns and unexpected events in the environment.

• Khaksar, Weria; Saplacan, Diana; Bygrave, Lee Andrew & Tørresen, Jim (2025). Robotics in Elderly Healthcare: A Qualitative Analysis of 20 Recent European Research Projects. ACM Transactions on Human-Robot Interaction. 14(2). doi: 10.1145/3711936. Full text in Research Archive
• Baselizadeh, Adel; Lindblom, Diana Saplacan; Khaksar, Weria; Uddin, Md Zia & Tørresen, Jim (2024). Comparative Analysis of Vision-Based Sensors for Human Monitoring in Care Robots: Exploring the Utility-Privacy Trade-off. In Sgorbissa, Antonio & Wei, Lei (Ed.), 2024 33rd IEEE International Conference on Robot and Human Interactive Communication (ROMAN). IEEE (Institute of Electrical and Electronics Engineers). ISSN 9798350375022. p. 1794–1801. doi: 10.1109/ro-man60168.2024.10731223. Full text in Research Archive
• Watanabe, Shin; Horn, Geir; Tørresen, Jim & Ellefsen, Kai Olav (2024). Improving Robot Skills by Integrating Task and Motion Planning with Learning from Demonstration. In Capi, Genci (Eds.), 2024 6th International Conference on Control and Robotics. IEEE (Institute of Electrical and Electronics Engineers). ISSN 9798331518158. doi: 10.1109/iccr64365.2024.10927043. Full text in Research Archive
• Wallace, Benedikte; Nymoen, Kristian; Tørresen, Jim & Martin, Charles Patrick (2024). Breaking from realism: exploring the potential of glitch in AI-generated dance. Digital Creativity. ISSN 1462-6268. 35, p. 125–142. doi: 10.1080/14626268.2024.2327006. Full text in Research Archive Show summary

  What role does deviation from realism play in the potential for generative artificial intelligence (AI) as a creative tool? A deep case-study was performed to explore interactions with AI-generated dance sequences as an inspiration source in dance composition and improvization. We present a simple interface created in collaboration with an experienced dancer for browsing AI-generated dance. By including glitches, the physics-breaking mistakes often encountered in AI-generated artefacts, we examine their affordances and possible use cases through sessions with the dancer. Through a process of reflexive thematic analysis, we identified that generative AI can engage a dancer through surprise, inspiring a transformation from abstract to physical movements. Our work challenges existing notions of the importance of realism in dance generation models, exemplifies the importance of close collaboration with practitioners in evaluating AI-generated artefacts and proposes glitch as a potential use case for dance ideation as it encourages dancers to embody unfamiliar movement qualities and break from ingrained patterns.

• Hansen, Tom Frode; Liu, Zhongqiang & Tørresen, Jim (2024). Predicting rock type from MWD tunnel data using a reproducible ML-modelling process. Tunnelling and Underground Space Technology. ISSN 0886-7798. 152. doi: 10.1016/j.tust.2024.105843. Full text in Research Archive Show summary

  Despite the increasing global usage of Measure While Drilling (MWD) data in tunnel projects, the application of machine learning (ML) techniques for real-time rock type prediction still needs to be explored. This paper introduces a novel ML approach to predict rock types in advance of the tunnel face using MWD data. Drawing on a diverse dataset of 4986 samples from 15 Norwegian tunnels, this study employed a pipeline including the LightGBM machine learning algorithm to forecast rock types 3–6 m ahead of excavation, achieving a balanced accuracy of 0.96 for six primary rock types using all 48 MWD-features. A more challenging label configuration with ten classes of near similar rock types, aimed to test the outer boundaries of model performance, achieved a score of 0.87. Performance in geologic transition zones is compared to regular zones. Notably, this capability facilitates proactive logistics for excavated rock material reuse and rock engineering strategies. Data leakage and reproducibility challenges in ML-based research are addressed in a step-by-step approach, and comparisons are drawn between digital and conventional scientific experimentation.

• Baselizadeh, Adel; Uddin, Md Zia; Khaksar, Weria; Saplacan, Diana & Tørresen, Jim (2024). PriMA-Care: Privacy-Preserving Multi-modal Dataset for Human Activity Recognition in Care Robots. In Grollman, Dan & Broadbent, Elizabeth (Ed.), HRI '24: Companion of the 2024 ACM/IEEE International Conference on Human-Robot Interaction. Association for Computing Machinery (ACM). ISSN 9798400703232. p. 233–237. doi: 10.1145/3610978.3640701. Full text in Research Archive Show summary

  In the field of robotics, caregiving robots and personal assistants are assuming an increasingly prominent role, directly impacting human lives. Especially in healthcare domains, these systems are starting to provide continuous 24/7 care by monitoring patients and delivering real-time insights into their activities. The effective deployment of future robots relies on equipping them with sophisticated Human Activity Recognition (HAR) algorithms. Many HAR algorithms are based on Artificial Intelligence (AI) and Machine Learning (ML) models. The development of these models necessitates suitable datasets. This paper introduces a Privacy-preserving Multimodal dataset for HAR in the context of Human-Robot Interaction (HRI) for Care robots (PriMA-Care). Tailored for care robots, PriMA-Care includes 27 diverse user activities, spanning daily tasks to physical HRI, with data from 10 privacy-preserving sensors and 17 participants. PriMA-Care addresses critical gaps in existing datasets, offering a suitable resource for HAR research in care robots.

• Saplacan, Diana; Pajalic, Zada & Tørresen, Jim (2024). Should Social and Assistive Robots Integrated within Home- and Healthcare Services Be Universally Designed? In Barfield, Woodrow; Weng, Yueh-Hsuan & Pagallo, Ugo (Ed.), The Cambridge Handbook of the Law, Policy, and Regulation for Human–Robot Interaction. Cambridge University Press. ISSN 9781009386708. p. 252–287. doi: 10.1017/9781009386708.018. Full text in Research Archive
• Otterdijk, Maria Theodorus Henricus van; Neggers, Margot; Tørresen, Jim & Barakova, Emilia (2024). A Study on Congruency in Expressive Robot Movement While Delivering Food to Seniors. In Zhang, Houxiang & Shi, Qing (Ed.), Proceedings of the 2024 IEEE International Conference on Real-time Computing and Robotics (RCAR). IEEE (Institute of Electrical and Electronics Engineers). ISSN 9798350372601. p. 431–436. doi: 10.1109/RCAR61438.2024.10670931. Full text in Research Archive
• Carlsen, Per Antoine; Taylor, Angelique M.; Chan, Darren M; Uddin, Md Zia; Riek, Laurel & Tørresen, Jim (2024). Real-Time Person Re-Identification to Improve Human-Robot Interaction. In Zhang, Houxiang & Shi, Qing (Ed.), Proceedings of the 2024 IEEE International Conference on Real-time Computing and Robotics (RCAR). IEEE (Institute of Electrical and Electronics Engineers). ISSN 9798350372601. p. 425–430. doi: 10.1109/RCAR61438.2024.10671043. Full text in Research Archive
• Otterdijk, Marieke van; Laeng, Bruno; Saplacan, Diana & Tørresen, Jim (2024). The Effect of Expressive Robot Behavior on Users’ Mental Effort: A Pupillometry Study. IEEE Transactions on Cognitive and Developmental Systems. ISSN 2379-8920. 16(2), p. 474–484. doi: 10.1109/TCDS.2024.3352893. Full text in Research Archive Show summary

  Robots are becoming part of our social landscape. Social interaction with humans must be efficient and intuitive to understand because nonverbal cues make social interactions between humans and robots more efficient. This study measures mental effort to investigate factors influencing the intuitive understanding of expressive nonverbal robot motions. Using an eye tracker to measure pupil response and gaze, fifty participants were asked to watch eighteen short video clips featuring three different types of robots performing expressive robot behaviors. Our findings indicate that the appearance of the robot, the viewing angle, and the expression shown by the robot all influence the cognitive load. Therefore, they may affect the intuitive understanding of expressive robot behavior. Furthermore, we found differences in the fixation time for different features of the various robots. With these insights, we identified possible improvement directions for making interactions between humans and robots more efficient and intuitive.

• Saplacan, Diana; Otterdijk, Marieke van & Tørresen, Jim (2024). Qualitative Observational Video-Based Study on Perceived Privacy in Social Robots’ Based on Robots Appearances. Advanced Robotics and its Social Impacts (ARSO). ISSN 2162-7568. doi: 10.1109/ARSO60199.2024.10557941. Full text in Research Archive Show summary

  Privacy has recently got attention, especially since the introduction of the General Data Protection Regulation (GDPR) in Europe, the new Artificial Intelligence Act (AIA) and the new Machinery Regulation. Privacy can be defined as someone’s right to keep personal matters, including personal life, personal information, or relationships, to themselves. A social robot’s appearance (=the combination of embodiment and motion) may contribute to how human users perceive them, including how these robots are perceived in relation to privacy. If these robots are part of certain services such as home- or healthcare, these may also have consequences on how these services are perceived. This study aims at showcasing the users’ perception of privacy based on the perceived robot’s appearance. Three social robots were chosen for this purpose: PLEO (with a zoomorphic appearance), Pepper (with a childlike anthropomorphic appearance), and TIAGo (with a mechanical and asymmetrical appearance). The data was collected through an in-lab observational video-based study from 50 participants with very limited- or no experience with robots. Our findings show that PLEO was perceived as preserving most of the users' privacy, while Pepper was perceived as more privacy-invasive than PLEO but less than TIAGo. TIAGo was perceived as hard to interpret in terms of privacy. Our findings also point out that designing robots with a cute appearance, such as PLEO, may contribute to participants trusting the robot more and thus being willing to share their data. The paper provides a list of characteristics that participants associated with a social robot as preserving or not their privacy. Further, the paper discusses the appearance of these social robots in terms of “cuteness” as a dark pattern in the design of social robots that may lead to data myopia, but also the possible consequences this may have, for vulnerable users, while trying to design more inclusive robots.

• Lømo, Tobias; Tørresen, Jim; Kolberg, Mariana & Maffei, Renan (2024). Multi Map Visual Localization for Unmanned Aerial Vehicles. IEEE Robotics and Automation Letters. ISSN 2377-3766. 10(2). doi: 10.1109/LRA.2024.3518071. Full text in Research Archive Show summary

  Localization has long been an essential area of research within robotics. The popularity of using Unmanned Aerial Vehicles (UAVs) to solve different tasks has increased and is expected to continue. Developing a robust complementary system to the Global Navigation Satellite Systems (GNSS) used today has been researched, and visual localization using cameras and satellite images is a popular choice to use. One of the challenges with using satellite images is that different images over the same area can impact the system's performance. This article proposes a novel approach called Multi Map Visual Localization (MMVL), a method to use multiple satellite images simultaneously, which is combined using a weighted average of probability maps. The proposal uses a convolutional neural network (CNN) with a caching strategy together with Monto Carlo Localization (MCL). MMVL achieves excellent robustness compared to other approaches and manages to estimate the correct location on all test flights. At the same time, using multiple satellite images does not significantly impact accuracy and computation time.

• Thorsen, Ola; Esema, Emmanuel Joseph; Hemaz, Said; Ellefsen, Kai Olav; Herrebrøden, Henrik & Arnim, Hugh Alexander von [Show all 7 contributors for this article] (2024). Can machine learning help reveal the competitive advantage of elite beach volleyball players? In Westphal, Florian; Peretz-Andersson, Einav; Riveiro, Maria; Bach, Kerstin & Heintz, Fredrik (Ed.), 14th Scandinavian Conference on Artificial Intelligence SCAI 2024, June 10-11, 2024, Jönköping, Sweden. Swedish Artificial Intelligence Society. ISSN 9789180757096. p. 57–66. doi: 10.3384/ecp208007. Full text in Research Archive Show summary

  As the world of competitive sports increasingly embraces data-driven techniques, our research explores the potential of machine learning in distinguishing elite from semi-elite beach volleyball players. This study is motivated by the need to understand the subtle yet crucial differences in player movements that contribute to high-level performance in beach volleyball. Utilizing advanced machine learning techniques, we analyzed specific movement patterns of the motion of the torso during spikes, captured through vest-mounted accelerometers. Our approach offers novel insights into the nuanced dynamics of elite play, revealing that certain movement patterns are distinctly characteristic of higher skill levels. One of our key contributions is the ability to classify spiking movements at different skill levels with an accuracy rate as high as 87%. This current research provides a foundation of what separates elite players from their semi-elite counterparts.

• Sandanger, Tonje Viddal; Sivachandran, Annica; Eriksen, Roman Stenseth; Baselizadeh, Adel; Otterdijk, Maria van & Kwak, Dongho [Show all 7 contributors for this article] (2024). Users’ perception and dynamic motion velocity matching in human-robot approach scenarios. In Harada, Kensuke; Demircan, Emel & Weiss, Astrid (Ed.), Proceedings of 2024 IEEE International Conference on Advanced Robotics and Its Social Impacts (ARSO 2024). IEEE (Institute of Electrical and Electronics Engineers). ISSN 9798350344646. p. 67–73. doi: 10.1109/ARSO60199.2024.10557976. Full text in Research Archive
• Kwak, Dongho; Baselizadeh, Adel; Otterdijk, Marieke van; Saplacan, Diana & Tørresen, Jim (2024). "Noisy" Matching of Motion Velocity of an Assistive Robot to the Users' Walking Velocity. ACM/IEEE International Conference on Human-Robot Interaction (HRI). ISSN 2167-2121. doi: 10.1145/3610978.3640546. Full text in Research Archive Show summary

  This study investigates the impact of dynamic matching of robot motion velocity to users' walking velocity in a human-robot approach scenario on three categories: perceived comfort, interactivity, and naturalness. Considering age diversity, participants were divided into two age groups. Young participants tended to rate higher for all three categories when the robot's approaching velocity was dynamically matched. In contrast, elderly participants preferred a steady and slow robot approach for comfort and predictability. These findings contribute to the ongoing effort to design assistive robots that effectively cater to diverse user groups, ultimately enhancing user satisfaction and acceptance. Taken together, this study highlights the importance of dynamically tailoring robot behaviors based on user demographics for positive Human-Robot Interactions.

• Otterdijk, Marieke van; Saplacan, Diana; Laeng, Bruno & Tørresen, Jim (2024). An Exploratory Study on People's Intuitive Understanding of Expressive Robot Behavior. ACM/IEEE International Conference on Human-Robot Interaction (HRI). ISSN 2167-2121. p. 1072–1076. doi: 10.1145/3610978.3640568. Full text in Research Archive Show summary

  Robots are anticipated to communicate with humans in our social context. Using nonverbal communication by robots increases communication comprehension because humans intuitively understand such behavior due to their experience of human-human interaction. Thus, this behavior is suitable for studying what makes robot motion intuitive to understand in human-robot interaction (HRI). We study how robot features help humans intuitively grasp expressive robot gestures, concentrating on what makes behavior easy to interpret. After watching eighteen nine-second videos of three robot kinds demonstrating expressive robot actions, 50 participants completed an open-ended survey. Our findings highlight the inputs, mediating factors, and outputs that users reported based on observing examples of expressive robot behavior. These insights are a starting point for analyzing robot behavior from the perspective of intuition and provide a foundation for a theoretical framework for intuition in HRI.

• Wallace, Benedikte; Otterdijk, Marieke van; Zhang, Yuchong; Rajabi, Nona; Marin-Bucio, Diego & Kragic, Danica [Show all 7 contributors for this article] (2024). Imitation or Innovation? Translating Features of Expressive Motion from Humans to Robots, HAI '24: Proceedings of the 12th International Conference on Human-Agent Interaction. Association for Computing Machinery (ACM). ISSN 9798400711787. p. 296–304. doi: 10.1145/3687272.3688302. Full text in Research Archive

View all works in NVA

• Watanabe, Shin; Horn, Geir; Tørresen, Jim & Ellefsen, Kai Olav (2025). 2025 IEEE 21st International Conference on Automation Science and Engineering (CASE). Institute of Electrical and Electronics Engineers (IEEE). ISBN 9798331522476. doi: https:/ieeexplore.ieee.org/xpl/conhome/11163731/proceeding. Full text in Research Archive

View all works in NVA

• Tørresen, Jim (2026). Tutorial: Ethics in Embedded Systems and Robots. Full text in Research Archive
• Tørresen, Jim (2026). Embedding Intelligence in Robots Through Interdisciplinary Collaboration. Full text in Research Archive
• Fossum, Mariann; Hansen, Michael Rygaard; Thygesen, Elin; Tørresen, Jim; Melkas, Helinä & Losada-Floriano, Diego (2026). SAKO project seminar. Full text in Research Archive Show summary

  Seminar June 2026 for all participants involved in the project with invited presenters.

• Tørresen, Jim (2026). 2nd Workshop on Robot Ethics – Ethical, Legal and User Perspectives in Robotics and Automation (WOROBET). Full text in Research Archive
• Lindblom, Diana Saplacan; Kuoppamäki, Sanna; Bodenhagen, Leon & Tørresen, Jim (2026). Presentation during Workshop at NordiCHI 2026 at the 1st Workshop on Robots as Welfare Technologies and Actors within Future Home- and Healthcare Services: Co-Designing an Integration Model through Story-Dialogue Method. Full text in Research Archive
• Sikora, Claudia Magdalena; Lindblom, Diana Saplacan & Tørresen, Jim (2026). Presentation of accepted paper: “There’s something human about it!” - Exploring users’ perception of nonverbal communication of robotic furniture. doi: https:/icsr2026.uk/programme/. Full text in Research Archive Show summary

  Non-anthropomorphic robots are integrating into our homes. People attribute human-like characteristics to these robots, although they do not resemble humans. This raises the question of whether similar processes would occur when people interact with robot furniture. Thus, the presented exploratory pilot study investigates how users intuitively interpret and perceive the non verbal communication of such robots. In an instrumental case study, we had eleven participants interacting with the T-ABLE robot through expressive behavior. Participants were interviewed about their experience. Our findings indicate that the participants use their associations between behavior and prior to similar experiences to attribute personality to the robot table. These findings provide insights for theorizing the intuitive comprehension of the behavior of non-anthropomorphic robots.

• Lindblom, Diana Saplacan; Kuoppamäki, Sanna; Bodehagen, Leon & Tørresen, Jim (2026). Webpage for 1st Workshop on Robots as Welfare Technologies and Actors within Future Home- and Healthcare Services: Co-Designing an Integration Model through Story-Dialogue Method ROBOWELL. doi: https:/sites.google.com/view/robowell-nordichi26/home?pli=1&authuser=1. Full text in Research Archive
• Tørresen, Jim; Lindblom, Diana Saplacan; Prestes, Edson; Schweighofer, Erich; Caleb-Solly, Praminda & Weng, Yueh hsuan (2026). 2nd ICRA 2026 Workshop on Robot Ethics Ethical, Legal and User Perspectives in Robotics & Automation WOROBET. Full text in Research Archive
• Lindblom, Diana Saplacan; Irai, Hamid Reza; Strøm, Benedicte Sørensen & Tørresen, Jim (2026). Presentation of ROBOWELL project for Elderly Facility Center. Full text in Research Archive
• Lindblom, Diana Saplacan; Advocaat, Johnny; Tørresen, Jim & Sørensen, Linda (2026). Presentation of ROBOWELL project for Oslo Municipality. Full text in Research Archive
• Lindblom, Diana Saplacan; Sørensen, Linda; Tørresen, Jim; Bodenhagen, Leon & Palinko, Oskar (2026). Presentation of ROBOWELL project for Sunnaas Rehabilitation Hospital. Full text in Research Archive
• Tørresen, Jim (2025). Presentation as a panelist at a panel on global perspectives on AI development and policy. Full text in Research Archive
• Tørresen, Jim (2025). Industrial AI Center Partner Presentation. Full text in Research Archive
• Tørresen, Jim (2025). How should Robot Assistants behave? Full text in Research Archive
• Tørresen, Jim (2025). From Adaptation of the Robot Body and Brain Using Rapid-Prototyping to Human–Robot Interaction with TIAGo. Full text in Research Archive
• Melteig, Elina; Lintvedt, Mona Naomi & Tørresen, Jim (2025). Selv luftmåleren kan avsløre en hjemmefest. Derfor er roboter i hjemmet perfekte verktøy for overvåkere. - Titan.uio.no. [Internet]. Titan.uio.no. doi: https:/www.titan.uio.no/teknologi/2025/roboter-i-hjemmet-utfordrer-personvernet.html. Full text in Research Archive Show summary

  – Uavhengig av hva som er intensjonen, får man mindre privatliv med roboter i hjemmet, sier forsker. Men det finnes noen råd for bedre personvern. 

• Lindblom, Diana Saplacan; Kuoppamäki, Sanna; Bodenhagen, Leon; Tørresen, Jim; Baselizadeh, Adel & Palinko, Oskar (2025). Organization of meeting for ROBOWELL project. Full text in Research Archive
• Guo, Jinyue; Tørresen, Jim & Jensenius, Alexander Refsum (2025). Cross-modal Analysis of Spatial-Temporal Auditory Stimuli and Human Micromotion when Standing Still in Indoor Environments (poster). doi: 10.5281/zenodo.17502603. Full text in Research Archive
• Tørresen, Jim; Prestes, Edson; Caleb-Solly, Praminda; Weng, Yueh Hsuan & Watanabe, Shin (2025). CASE 2025 Workshop on Ethical Considerations in Robotics and Automation (WECRA). doi: https:/sites.google.com/view/robot-ethics-case2025-workshop. Full text in Research Archive
• Mahler, Tobias & Tørresen, Jim (2025). VIROS in a Nutshell. doi: https:/www.jus.uio.no/ifp/english/research/projects/nrccl/viros/events/seminars/vulnerability-in-the-robot-society-final-conferenc.html. Full text in Research Archive
• Tørresen, Jim (2025). Keynote: Techno-Ethical Considerations when Applying Machine Learning in Real-world Systems. doi: https:/www.icmlt.org/2025.html. Full text in Research Archive
• Tørresen, Jim (2025). Invited talk: Intelligent Robotics in Healthcare. Full text in Research Archive
• Tørresen, Jim (2025). Guest lecture: Intelligent Robotics in (Home) Healthcare. Full text in Research Archive
• Tørresen, Jim (2025). Invited talk: When will a robot treat us instead of a medical doctor? doi: https:/www.uio.no/english/research/strategic-research-areas/life-science/norway-life-science-conference/. Full text in Research Archive
• Nergård, Katrine Linnea; Ellefsen, Kai Olav & Tørresen, Jim (2025). Fast or Slow: Adaptive Decision Maiking in Reinfocement Learning with Pre-Trained LLMs. Full text in Research Archive
• Santos, Leticia dos; Rosinski, Marcelo & Tørresen, Jim (2025). Robô autônomo pode auxiliar nas tarefas do dia a dia. [Internet]. Jornal da Universidade UFRGS. Full text in Research Archive
• Tapus, Adriana; Zhegong, Shangguan; Tørresen, Jim & Søraa, Roger Andre (2024). IEEE RO-MAN 2024 Workshop on Ethics Challenges in Socially Assistive Robots and Agents: Legality, Value Orientation, and Future Design for Human-Robot Interaction (HRI). doi: https:/perso.ensta-paris.fr/~shangguan/Ro-manWS/Home.html. Full text in Research Archive
• Tørresen, Jim; Prestes, Edson; Caleb-Solly, Praminda & Weng, Yueh Hsuan (2024). IEEE/RSJ IROS 2024 workshop Ethical, Legal and User Perspectives on Assisting Robots and Systems (WELUPARS). doi: https:/sites.google.com/view/iros-2024-robot-ethics?pli=1&authuser=1. Full text in Research Archive
• Tørresen, Jim (2024). Invited talk: Intelligent robots - the future of healthcare? Full text in Research Archive
• Tørresen, Jim (2024). Invited talk: Ethics Integrated in Human-Robot Interaction Research. doi: https:/www.roboethics.design. Full text in Research Archive
• Tørresen, Jim (2024). Ethical, Legal and Technical Challenges and Considerations. Full text in Research Archive
• Tørresen, Jim (2024). Invitert foredrag: Vil vi ha roboter til å hjelpe oss når vi trenger hjelp? Full text in Research Archive
• Tørresen, Jim (2024). Kunstig intelligens og forskningsetiske vurderinger. Full text in Research Archive
• Tørresen, Jim (2024). Tutorial: AI Ethical challenges and considerations. Full text in Research Archive
• Tørresen, Jim; Baselizadeh, Adel; Mahler, Tobias & Bygrave, Lee Andrew (2024). Machine Performance Tradeoffs to Ethical and Legal Perspectives. Full text in Research Archive
• Tørresen, Jim (2024). Ethical considerations for robots supporting independent living of older people. Full text in Research Archive
• Tørresen, Jim (2024). AI Ethics – Challenges and Opportunities – Keynote talk. Full text in Research Archive
• Tørresen, Jim (2024). Ethical, Legal and User Perspectives on Research in Assistive Robots. Full text in Research Archive
• Tørresen, Jim (2024). How human-like and interactive would we want a robot assistant to be? Full text in Research Archive
• Tørresen, Jim (2024). From Adaptation of the Robot Body and Control Using Rapid-Prototyping to Human–Robot Interaction with TIAGo. Full text in Research Archive
• Shi, Zhongzhi; Tørresen, Jim & Yang, Shenxiang (2024). Preface. IFIP Advances in Information and Communication Technology. ISSN 1868-4238. 704, p. v–v. Full text in Research Archive
• Tørresen, Jim (2024). Ethical and Regulatory Perspectives of Robotics and Automation. Full text in Research Archive
• Nishimura, Markus Yuki; Matsumoto, Kohei & Tørresen, Jim (2026). Diffusion-Based Reinforcement Learning for Multi-Robot Social Navigation. Universitetet i Oslo. Full text in Research Archive Show summary

  --- Abstract and files will be published: 2027-05-15

• Moe, Håkon Øyehaug; Tørresen, Jim & Iwashita, Yumi (2025). Disparity Estimation from JPEG-Compressed Lunar Stereo Pairs using Deep Stereo Matching and Conditional Diffusion Models. Universitetet i Oslo. Full text in Research Archive Show summary

  Accurate depth mapping is crucial for creating 3D lunar surface models, essential for future exploration and navigation missions. This thesis addresses challenges in generating precise depth estimates from JPEG-compressed stereo images acquired by Japan’s Kaguya lunar orbiter, which traditional algorithms struggle to handle effectively. To overcome these limitations, two deep learning methods were developed and tested. The first method uses a modified version of the IGEV++ stereo matching network, renowned for iteratively improving depth estimation accuracy. The second approach employs Palette, a conditional diffusion model adapted specifically to handle compressed stereo images. A preprocessing technique, which trains models to predict residual corrections rather than full depth values, further simplifies and enhances learning. Results demonstrate both models significantly outperform traditional approaches. Particularly, the Palette diffusion model achieves superior accuracy, highlighting the advantages of diffusion-based methods for reliably extracting detailed depth information from compressed lunar imagery. These advancements offer valuable insights for future lunar mission planning and satellite image processing strategies.

• Thorsen, Ola; Tørresen, Jim; Vellasco, Marley Maria Bernardes Rebuzzi & Thimoteo, Lucas Martins (2025). CNN Ensemble Multiclass Alzheimer’s Disease Stage Classification with Human Interpretable Results: A Single Modality Approach. Universitetet i Oslo. Full text in Research Archive Show summary

  In an age defined by the blistering pace of technological innovation, Alzheimer’s Disease remains a bastion of unsolved medical mystery. With cutting-edge artificial intelligence tools in hand, researchers worldwide are making herculean efforts, chipping away at this most stubborn of problems. Intending to contribute to these efforts, in this thesis, we explore the utilization of an ensemble of multiple Convolutional Neural Networks to classify the four disease stages: Cognitive Normal, Stable Mild Cognitive Impairment, Progressive Mild Cognitive Impairment, and Alzheimer’s Disease. By relying on only a single data modality, we maximize the size of our usable dataset. In addition, by performing a feature selection process tailored to each ensemble, we reduce the training data volume while increasing the human-centered interpretability of our ensemble model predictions. We propose and test three different EfficientNetV2- based ensemble architectures, achieving a macro average Area Under the Curves of 0.712, 0,724, and 0.735, respectively, on the four-class classification problem. Our best performing ensemble, aptly named the 3&2-Class ensemble, achieves a recall score of 0.419 on the Alzheimer’s Disease class compared to the baseline of 0.239. Furthermore, using the state-of-the- art post-hoc interpretability algorithm Shapley Additive Explanations, we incorporate Shapley values into our proposed human-centered explainability methods that allow for both local and global explanations. Together, these form a robust and interpretable framework for early detection and disease stage classification of Alzheimer’s disease.

• Bjerkan, Mons Eirik; Watanabe, Shin & Tørresen, Jim (2025). Offline Reinforcement Learning through MPC Bootstrapping. Universitetet i Oslo. Full text in Research Archive Show summary

  Efficient, safe, and robust learning for robotic tasks in the real world is more relevant than ever, with new robots or software improvements emerging rapidly. Accordingly, the popularity of \gls{rl}, and offline \gls{rl} has also massively increased over the last years. There are, however, issues training robots for real-life tasks using \gls{rl} with concerns of safety and the amount of data needed. Offline reinforcement learning is a paradigm with the promise of alleviating some of these challenges. Control algorithms such as \gls{mpc} also aim to solve these complex issues. This approach of using non-learning-based control algorithms has long been the most dominant paradigm within control of complex systems. And has only become more relevant with the emergence of real-time optimization, such as real-time \gls{mpc}. To leverage the best of both worlds, this paper proposes integrating offline \gls{rl} with \gls{mpc} by training an offline \gls{rl} policy solely on pre-collected \gls{mpc} trajectories, thereby effectively bootstrapping the \gls{rl} policy training process. The algorithm aims to avoid excessive exploration for faster convergence towards stronger solutions by leveraging the already provided solutions from the \gls{mpc}, while exploiting the characteristics of \gls{rl} to make up for any weaknesses in the \gls{mpc} solutions to make a robust, safe, and data-efficient solution for task-specific robot training. The proposed method is evaluated in a simulated environment and tested on a quadruped robot navigating through a hilly terrain. Testing shows that the proposed method does not measure up to the performance of pure \gls{mpc}. However, the results indicate that the proposed method can be applied to learn how to solve simpler control tasks.

• Sandanger, Tonje Viddal; Tørresen, Jim; Wiig, Ola; Homlong, Eirik Gromholt & Kumar, Rahul Prasanna (2025). Enhancing Gait Analysis through eXplainable AI. Universitetet i Oslo. Full text in Research Archive Show summary

  Etter kvart som kunstig intelligente (KI) system vert integrert i aukande grad i helsevesenet, er behovet for transparente, pålitelege og klinisk meiningsfulle verktøy viktigare en nokon gang. Denne avhandlinga undersøkjer bruken av maskinlæring og forklarbar KI i klinisk ganganalyse for barn med Celebral Parese (CP). Ved å utnytte rørsle-data frå røynda, utforskar denne studia bruken av djupe-lærings modeller som CNN-LSTM, RNN-LSTM, og ST-GCN, for å sjå korleis desse samanliknast med enklare modellar som RFar. Desse modellane vart trent til å klassifisere Gross Motor Function Classification System (GMFCS) nivå hos barn med CP. Ettersom behovet for pålitelege og forklarbare applikasjonar er avgjerande i helsevesenet, vart SHAP brukt til å lage forklaringer av modellane sine predikasjonar, noko som mogleggjer eit transparent og tolkbart system. Resultata vise at meir komplekse modellar, som ST-GCNane, vart utkonkurrert til fordel for enklare modellar under dataavgrensande forhald, og at dei best ytande modellane vis signifikante resultat for forklaringer med fokus på klinisk relevante funksjonar som korresponderer med det rørte eller kontrarørte beinet. Dette arbeidet framheva óg røynda av nøye modellvalg når ein designar forklarbare KI system, ettersom at SHAP er svært modellavhengig. Det vart derimot funne at når ein brukar ensamble-metodar, vart den totale betydninga av funksjonane fokusert på klinisk relevante plan og ledd, noko som framhev den lovande framtida til forklarbare system. Nokon nye bidrag frå arbeidet er dei tidsmessig justerte SHAP-varmekarta, overlappa gangsyklusen, som mogeleggjer for klinikarar å spore modell- slutningar til spesifikke biomekaniske hendingar. Sjølv om arbeidet bidrar til den aukande integrasjonen av KI-system i helsevesenet, står det att å validere den kliniske nytteverda gjennom direkte tilbakemeldingar frå klinikarar. Framtidig arbeid burde involvere bruker-sentrerte evalueringar, som integrerer domeneekspertar som gir ytterlegare innsikt i tolkbarheiten, relevansen og brukarvenlegheita til forklaringane. Ved å samkøyre denne forskinga med klinisk praksis legg den grunnlaget for bruk av forklarbar KI verktøy i klinisk ganganalyseapplikasjonar, ved å fokusere på openheit og forklarbarheit, samtidig som etiske praksisar vert overheldt.

• Lindblom, Diana Saplacan; Tørresen, Jim & Hakimi, Nina (2024). Dynamic Dimensions of Safety - How robot height and velocity affect human-robot interaction: An explorative study on the concept of perceived safety. University of Oslo. Full text in Research Archive
• Schulz, Trenton; Lindblom, Diana Saplacan; Tørresen, Jim & Badescu, Claudia Andreea (2024). Teach Me How To Dance: An exploration of motion translation methods using the ZTL framework for use with children with autism in educational settings. Universitetet i Oslo. Full text in Research Archive
• Lømo, Tobias; Maffei, Renan; Kolberg, Mariana & Tørresen, Jim (2024). Visual localization of unmanned aerial vehicles using multiple satellite maps. Universitetet i Oslo. Full text in Research Archive
• Chen, BiHui; Tørresen, Jim & Sanfilippo, Filippo (2024). Simulation-to-real-world reinforcement learning with the PAL TIAGo robot. Universitetet i Oslo. Full text in Research Archive Show summary

  Reinforcement learning(RL) has been popular in the robotics field due to its ability to adapt to complex, dynamic situations. Conducting RL in the robotics field faces several challenges, such as sample efficiency problems, where the RL algorithm often requires a large amount of data sample to train RL agents to learn a desired behavior. This thesis proposes an RL framework that leverages the use of simulated data by training RL agents in the simulated environment. The model-free reinforcement learning algorithm and Proximal policy optimization algorithm were used in this thesis. Domain randomization techniques were applied to develop multiple simulated environments where different parameters were chosen and randomized. Subsequently, we evaluated the performance of each agent from these training environments on the physical TIAGo robot and compared their difference in performance. Furthermore, we used the Gazebo simulator to integrate with ROS to simulate the training environment. We also provide a dense reward design that guides agents' training to learn a desired behavior, which is described as learning to "reach," "aim," and "hold the gripper parallel" to reach a desired pre-grasp pose. The test conducted on the physical TIAGo robot shows that the learned pre-grasp pose can be utilized to successfully grasp a cube. The best performance agent, trained in the fully randomized environment, achieved an average success of 44% across five different tested cube placements when transferred to the physical TIAGo robot. In comparison, the agent trained in the fixed environment setup achieved an average success of 6% when transferred to the physical TIAGo robot. The results suggest that training in a higher variability based on the training environment setup used in this thesis improves transferability when transferring the agent from the simulation to the real world in terms of a higher empirical success rate and more consistent performance.

• Sabamali, Melinda; Miura, Jun & Tørresen, Jim (2024). Confidence Estimation Using Verbal and Non-Verbal Features. Universitetet i Oslo. Full text in Research Archive
• Helgesen, Sander Elias Magnussen; Nakashima, Kazuto; Kurazume, Ryo & Tørresen, Jim (2024). Conditional Denoising Diffusion Probabilistic Models for Faster LiDAR Synthesis. Universitetet i Oslo. Full text in Research Archive Show summary

  3D LiDAR (Light Detection And Ranging) sensors are widely used in autonomous systems due to their accuracy and robustness to varying light and weather conditions. However, the sensor's resolution plays a great role in performance, and high-resolution sensors are costly. Existing approaches have shown promising results of using generative machine learning models, such as diffusion models, for upsampling low-resolution samples to high-resolution samples with high fidelity. However, these approaches are limited by slow inference speed due to the complexity of the models. Such limitations make these models fall short when it comes to real-time performance, making them unusable for real-world tasks such as autonomous navigation and human-robot interaction. This thesis introduces a novel approach based on conditional diffusion models to speed up sparse-to-dense upsampling of 3D LiDAR data while also improving sample generation fidelity. The approach builds on denoising diffusion probabilistic models (DDPMs) and conditions the input on an inpainting task, previously shown to perform well on image completion tasks. The A series of experiments is introduced to determine the effects and possible benefits of conditional inpainting. Furthermore, the effects of multi-objective learning are compared to single-objective learning. In addition, the approach is tested on various datasets to show its generalizing capabilities to various environments and sensor types. This thesis introduces the best model that strikes a balance between inference speed and generation fidelity. The conditional diffusion approach achieves a 39-times speedup compared to the best-performing baselines, with minimal performance sacrifice. Evaluation metrics scores such as IoU are shown to increase by 16% to 18% for the generated samples.

• Baselizadeh, Adel; Tørresen, Jim; Saplacan, Diana; Khaksar, Weria & Mahler, Tobias (2024). Advancing Safety and Privacy in Human-Robot Interaction: Exploring Learning-Based Control and Multi-Modal Sensing Approaches. Universitetet i Oslo. Full text in Research Archive Show summary

  As robots are imported from industrial settings to homes and healthcare environments, new risks emerge. Chief among these are concerns about user safety and privacy, which must be carefully considered during the development and integration of robots into daily life. Ensuring these aspects is essential for building trust between humans and robots, allowing them to function effectively and safely in close proximity to people. This thesis addresses these challenges by focusing on two key areas: improving safety through addressing the human-robot collision avoidance problem, and enhancing privacy through the use of privacy-preserving sensors in robots. In terms of safety, advanced control frameworks grounded in machine learning algorithms, control theory methods, and the integration of these two approaches have been developed to address the collision avoidance problem for robotic arms, ensuring safe operation without compromising efficiency. Regarding privacy, the thesis explores the use of privacy-preserving sensors such as thermal cameras and 3D lasers. These sensors offer effective monitoring while minimizing the collection of sensitive data, maintaining a balance between utility and privacy. By integrating these sensors with multimodal deep learning models, robots can recognize human activities and monitor users during human-robot interaction without invading personal privacy. This thesis contributes to safer and more private human-robot interactions, particularly in sensitive environments like healthcare, paving the way for more trusted and widely accepted robotic systems in daily life.

View all works in NVA