More Like this

1. Humans Need Augmented Feedback to Physically Track Non-Biological Robot Movements

   https://doi.org/10.1109/ICRA48891.2023.10161075  

   Edraki, Mahdiar ; Maurice, Pauline ; Sternad, Dagmar ( May 2023 , IEEE International Conference on Robotics and Automation)

   An important component for the effective collaboration of humans with robots is the compatibility of their movements, especially when humans physically collaborate with a robot partner. Following previous findings that humans interact more seamlessly with a robot that moves with humanlike or biological velocity profiles, this study examined whether humans can adapt to a robot that violates human signatures. The specific focus was on the role of extensive practice and realtime augmented feedback. Six groups of participants physically tracked a robot tracing an ellipse with profiles where velocity scaled with the curvature of the path in biological and nonbiological ways, while instructed to minimize the interaction force with the robot. Three of the 6 groups received real-time visual feedback about their force error. Results showed that with 3 daily practice sessions, when given feedback about their force errors, humans could decrease their interaction forces when the robot’s trajectory violated human-like velocity patterns. Conversely, when augmented feedback was not provided, there were no improvements despite this extensive practice. The biological profile showed no improvements, even with feedback, indicating that the (non-zero) force had already reached a floor level. These findings highlight the importance of biological robot trajectories and augmented feedback to guide humans to adapt to non-biological movements in physical human-robot interaction. These results have implications on various fields of robotics, such as surgical applications and collaborative robots for industry. 

   more » « less
2. Collaborative Research: Experiential Process Safety Training for Chemical Engineers, poster presentation and peer-reviewed proceedings, ASEE National Meeting, Tampa, FL.

   Anastasio, D ; Butler, B ; Burkey, D ; Cooper, M ; Bodnar, C ( January 2019 , ASEE Annual Conference proceedings)

   Process safety has become a critical component of chemical engineering education. However, students may find it difficult to fully understand the ramifications of decisions they make during classroom exercises due to their lack of real world experience. Use of an immersive digital environment where students could role play as chemical engineering employees making process safety decisions could be one method of achieving this goal. Through this experience, students could observe the outcomes of their decisions in a safe, controlled environment without the disastrous real-world consequences that could come from making a mistake. This digital environment could have further features, such as time constraints or interactions with other characters, to make the experience feel more authentic than an in-class discussion or case study. In order to evaluate the efficacy of such a virtual environment, a portion of this work centered around the creation of the Engineering Process Safety Research Instrument (EPSRI). The instrument asks participants to evaluate process safety dilemmas and rank a set of considerations based on how influential they were in their decision-making process. The instrument then classifies each decision based on the stages of Kohlberg’s moral development theory, ranging from pre-conventional (i.e. more self-centered) thinking to post-conventional (i.e. more global) thinking. This instrument will be used to assess how students’ thinking about process safety decisions changes as a result of engaging in the virtual safety decision making environment. This paper will summarize the progress since the project’s start in summer 2017,, highlighting the work completed in development and validation of the EPSRI. This process included content validation, think-aloud studies to improve clarity of the instrument, and factor analysis based on a large scale implementation at multiple universities. The paper will also discuss the development of the minimum viable product digital process safety experience, including establishment of learning outcomes and the mechanics that reinforce those outcomes. By presenting these findings, we intend to spread awareness of the EPSRI, which can evaluate the safety decisions of chemical engineering students while having the potential to launch discussions about safety and ethics in other engineering disciplines. We also hope that these results will provide educators with insights into how to translate educational objectives to elements of a digital learning environment through collaboration with digital media companies. 

   more » « less
3. Collaborative Research: Experiential Process Safety Training for Chemical Engineers

   Anastasio, D. D. ; Butler, B. L. ; Burkey, D. D. ; Cooper, M. ; Bodnar, C. A. ( June 2019 , 2019 ASEE Annual Conference & Exposition)

   Process safety has become a critical component of chemical engineering education. However, students may find it difficult to fully understand the ramifications of decisions they make during classroom exercises due to their lack of real world experience. Use of an immersive digital environment where students could role play as chemical engineering employees making process safety decisions could be one method of achieving this goal. Through this experience, students could observe the outcomes of their decisions in a safe, controlled environment without the disastrous real-world consequences that could come from making a mistake. This digital environment could have further features, such as time constraints or interactions with other characters, to make the experience feel more authentic than an in-class discussion or case study. In order to evaluate the efficacy of such a virtual environment, a portion of this work centered around the creation of the Engineering Process Safety Research Instrument (EPSRI). The instrument asks participants to evaluate process safety dilemmas and rank a set of considerations based on how influential they were in their decision-making process. The instrument then classifies each decision based on the stages of Kohlberg’s moral development theory, ranging from pre-conventional (i.e. more self-centered) thinking to post-conventional (i.e. more global) thinking. This instrument will be used to assess how students’ thinking about process safety decisions changes as a result of engaging in the virtual safety decision making environment. This paper will summarize the progress since the project’s start in summer 2017, highlighting the work completed in development and validation of the EPSRI. This process included content validation, think-aloud studies to improve clarity of the instrument, and factor analysis based on a large scale implementation at multiple universities. The paper will also discuss the development of the minimum viable product digital process safety experience, including establishment of learning outcomes and the mechanics that reinforce those outcomes. By presenting these findings, we intend to spread awareness of the EPSRI, which can evaluate the safety decisions of chemical engineering students while having the potential to launch discussions about safety and ethics in other engineering disciplines. We also hope that these results will provide educators with insights into how to translate educational objectives to elements of a digital learning environment through collaboration with digital media companies. 

   more » « less
4. Inferring Relational Potentials in Interacting Systems

   Comas, A. ; Du, Y. ; Fernandez Lopez, C. ; Ghimire, S. ; Sznaier, M. ; Tenenbaum, J. B. ; Camps, O. ( January 2023 , Proceedings of Machine Learning Research)

   Systems consisting of interacting agents are prevalent in the world, ranging from dynamical systems in physics to complex biological networks. To build systems which can interact robustly in the real world, it is thus important to be able to infer the precise interactions governing such systems. Existing approaches typically discover such interactions by explicitly modeling the feed-forward dynamics of the trajectories. In this work, we propose Neural Interaction Inference with Potentials (NIIP) as an alternative approach to discover such interactions that enables greater flexibility in trajectory modeling: it discovers a set of relational potentials, represented as energy functions, which when minimized reconstruct the original trajectory. NIIP assigns low energy to the subset of trajectories which respect the relational constraints observed. We illustrate that with these representations NIIP displays unique capabilities in test-time. First, it allows trajectory manipulation, such as interchanging interaction types across separately trained models, as well as trajectory forecasting. Additionally, it allows adding external hand-crafted potentials at test-time. Finally, NIIP enables the detection of out-of-distribution samples and anomalies without explicit training. 

   more » « less
5. Inferring Relational Potentials in Interacting Systems

   Comas, A. ; Du, Y. ; Fernandez Lopez, C. ; Ghimire, S. ; Sznaier, M. ; Tenenbaum, J.B. ; Camps, O. ( July 2023 , Proceedings of Machine Learning Research)

   Systems consisting of interacting agents are prevalent in the world, ranging from dynamical systems in physics to complex biological networks. To build systems which can interact robustly in the real world, it is thus important to be able to infer the precise interactions governing such systems. Existing approaches typically dis- cover such interactions by explicitly modeling the feed-forward dynamics of the trajectories. In this work, we propose Neural Interaction Inference with Potentials (NIIP) as an alternative approach to discover such interactions that enables greater flexibility in trajectory modeling: it discovers a set of relational potentials, represented as energy functions, which when minimized reconstruct the original trajectory. NIIP assigns low energy to the subset of trajectories which respect the relational constraints observed. We illustrate that with these representations NIIP displays unique capabilities in test-time. First, it allows trajectory manipulation, such as interchanging interaction types across separately trained models, as well as trajectory forecasting. Additionally, it allows adding external hand-crafted potentials at test-time. Finally, NIIP enables the detection of out-of-distribution samples and anomalies without explicit training. 

   more » « less