Organized Sessions

Emergence of “Cohabitant” Robots and Their ELSI Implications

In this session, we will examine ELSI based on interdisciplinary research methods for robots that we call “cohabitant robots,” which interact with people while building an emotional relationship with them, and which promote changes in human cognition and behavior.

While it has been pointed out that “cohabitant robots” have the potential to improve people’s wellbeing through improving mental health, maintaining cognitive abilities, and providing learning opportunities that are optimized for the individual, there are also concerns about the misuse of intimate personal data and the risk of inappropriate manipulation using emotional connections.

To control the risks of “cohabitant robots,” we will discuss these risks based on research using methods such as cognitive neuroscience, social psychology, and cultural anthropology. We will then explore the legal and technical measures that should be taken to reduce these risks from the perspectives of law and robotics under the framework of “Agile Governance.”

Topics of interest include (but are not limited to):

  1. The concept of “cohabitant robots” and the ELSI issues surrounding them.
  2. Research from the perspectives of cognitive neuroscience and social psychology on the impact of “cohabitant robots” on people.
  3. An anthropological exploration of the influence of “cohabitant robots” on human behaviors and interactions.
  4. Ethics of cohabitant robots: A study on the relationship between the framework for risk management of “cohabitant robots” (“Agile Governance”) and the ethics of robot engineers.

Organizers

  • Tatsuhiko Inatani (Kyoto University)
  • Minoru Asada (The University of Osaka)
  • Hiroko Kamide (Kyoto University)
  • Hirofumi Katsuno (Doshisha University)

Synthetic Data Generation for Reflective and Transparent Objects to Enhanced Robot Manipulation Performance

Recycling sorting is crucial in promoting sustainability and has long been recognized as a significant contribution to industrial processes. A wide variety of recyclable materials, such as plastic, glass, and metal, are routinely sorted. However, challenges persist/arise when handling reflective and transparent objects, such as metal cans and water bottles. These items present difficulties due to their complex optical properties and the shadows they cast, which can hinder accurate detection and classification.

In this paper, we propose a novel solution to address these challenges by focusing on the detection of transparent bottles and the accurate identification of reflective objects like metal cans. Our approach involves generating synthetic data of various recyclable objects to develop and enhance detection models specifically tailored for these challenging materials. By leveraging synthetic data, we can increase both the quantity and quality of training data, thereby improving the performance of object detection algorithms.

These 3D models accurately represent the characteristics and properties of real-world objects, ensuring that the synthetic data aligns with actual shapes, sizes, and textures. Our work improves the accuracy and efficiency of sorting systems, contributing to advancements in robotic picking and placing technologies. This work supports the evolving demands of modern industrial recycling processes, enabling more robust sorting capabilities that enhance overall sustainability efforts.

Topics of interest include (but are not limited to):

  1. Advanced industrial robots for future manufacturing
  2. Robotics for sustainability
  3. Service & assistance application

Organizers

  • Cheuk Tung Shadow YIU (The Hong Kong University of Science and Technology)
  • Hiu Ching LO (The Hong Kong University of Science and Technology)
  • Haolun HUANG (The Hong Kong University of Science and Technology)
  • Kam Tim WOO (The Hong Kong University of Science and Technology)