PDF version: 2026-fissure
Team(s) Loki & DEFROST (Inria Centre of the University of Lille – CRIStAL laboratory)
Level PhD
Duration 3 years
Advisor(s) Géry Casiez, Bruno Fruchard, Quentin Peyron [Contact advisor(s)]
This PhD is fully funded.
Abstract
While soft robotic manipulators are intrinsically safer than rigid manipulators, they may be perceived as threats due to diminished movement legibility, increased risks of movement singularities, and their animal-like appearances. This project leverages virtual reality (VR) setups and methods from speculative design to study perceived safety when users are immersed in interactive tasks with virtual realistic plant-like and animal-like soft robots. We posit the aesthetic of soft robots has a strong impact on this factor and it may have been overlooked.
Context & Goals
Robotic manipulators are used in various contexts to support human activities [10]. They are particularly useful in factories and construction lines to lift heavy objects or perform repetitive actions with precision that are physically and cognitively demanding for humans. Their size and strength poses challenges for the security of potential human co-workers, which has been extensively studied for rigid robots [7].
Soft robots emerged almost 30 years ago; they are often bio-inspired from animals or plants and are deformable by nature, which reduces the risks of damaging their environment upon impact [8]. They were recently identified as good candidates to create robots with minimal or positive ecological footprint. Their body naturally minimize the amount of material required to fabricate the structure, its inertia and therefore the energy required to put it in motion, and it can be fabricated using bio-sourced, plant-based materials [3]. Their design is actively discussed in the literature [4], but their interaction with and by human co-workers remain largely unstudied. While they may pose less physical threats, humans may not feel safe due to their non-anthropomorphic appearance (e.g., looking like animals workers are afraid of, see Figure 1), the uncommon materials they can be made of (silicone, bio-sourced
materials, etc), and their perceived behavior based on their movement legibility. Indeed, their motion and behavior is generally difficult to understand for a non-expert user. As a consequence, they are often used with middle/high level control algorithms coordinating the robot actuators for end-effector positioning tasks [1]. Furthermore, they can be subject to bulking, creating scenarios where the robot is either stuck or snaps suddenly with fast and dangerous movements.
This project challenges the fact that serial and parallel soft manipulators are safer than rigid robots by design, and makes the hypothesis their perceived safety may be lesser. We hypothesize users may feel threatened by their presence, particularly if they cannot easily predict their movements, partly due to the presence of singularities, and that their appearance triggers specific emotional responses (e.g., fear, stress, anxiety). To date, few studies investigated the impact of the movement legibility and appearance of rigid manipulators with regard to perceived safety [7], and even less for soft manipulators [9]. Studying these factors may imply to expose users to soft manipulator designs that vary in size, form factor, texture and degrees of freedom, which would require a significant amount of resources to build.
Instead, we propose to leverage virtual environments to simulate collaborative scenarios between human(s) and robot(s) [6] which would elicit the exact same physiological and emotional responses for humans than real use cases.
Through this PhD, our goals are to:
1. design an experimental protocol and setup using virtual environments to investigate the
perceived safety of soft robots in Human-Robot Interaction (HRI) tasks;
2. study the impact of singularities, movement legibility, and aesthetics of soft robots on
perceived safety;
3. draw design guidelines on the material and aesthetic of soft robots to maximize the
perceived safety and the overall user experience during HRI.
Research Questions
This research focuses on three primary factors: how the perceived and actual risks of singularities impacts perceived safety during HRI tasks, how the predictability and legibility of robot movements affect their perception [5], and how the robot appearance (i.e., realistic animal-like or plant-like appearances) impacts users emotional responses. We propose as a general methodology to focus on soft manipulators using 1 to 3 effectors, with 3~up to 18~degrees-of-freedom (DOFs), and a maximum amplitude of movement from 1 to 5 meters. We propose to investigate tasks inspired from manufacturing and supply chain contexts [6] in which a single human and a single robot manipulate virtual objects synchronously, for instance, to place them in targeted zones. We envision the robot movements to be predetermined animations to grab and release objects, in which we will purposefully introduce singularities depending on the experimental conditions. These animations will either be simulated through realistic physical simulation of soft structures using SOFA [2] and reproduced with models in VR to simplify integration if necessary, or produced with animation software (e.g., Blender) to simulate specific animal-like behaviors that may go beyond the capabilities of current physics-based simulators. We will adapt their speed and trajectories based on user movements to simulate interactivity.
We propose to focus on the following research questions:
1. How are singularities perceived by humans depending on the robot form factor and
the task?
2. How do the robot movement legibility and perceived behaviors impact perceived
safety?
3. Does the appearance of soft robots influence positively or negatively perceived safety?
Hosting Laboratory
The PhD candidate will join the Inria centre of the University of Lille and be part of both the LOKI and DEFROST research groups, respectively specialized in Human-Computer Interaction and Soft Robotics. Both teams are affiliated with the CRIStAL laboratory (UMR 9189) and includes professors and assistant professors of the University of Lille, as well as Inria researchers. Lille is at the northern tip of France and its metropolitan area, situated at the crossroads of northern continental Europe, is the 5th biggest in France. Both teams are dynamic and multicultural, with members coming from different countries (Germany, Canada, China, Iran, France, etc.) and communicating daily in English.
Candidate
The candidate must have (or be about to obtain) a Master’s degree or equivalent in Computer Science, Human-Computer Interaction, or (Soft) Robotics, and demonstrate a strong interest in research. We value overall creativity, independence, team spirit and communication skills. A good level of technical and scientific English is required.
We are looking for candidates with strong skills in:
• developing interactive applications, especially in VR (e.g., using Unity or Godot)
• knowledge in 3D modeling and animation (e.g., using Blender)
• knowledge in HCI and experimental design
Knowledge in the following topics are a plus:
• in soft robotics, e.g., types of robots and actuation
• in physics simulation, e.g., using SOFA
Application
To apply, send your resume and a cover letter by email to Géry Casiez (gery.casiez@univ-lille.fr), Bruno Fruchard (bruno.fruchard@inria.fr) and Quentin Peyron (quentin.peyron@inria.fr) with [Application: Studying Factors Involved in the Perceived Safety of Soft Robots] as object of the e-mail. In addition to what is generally expected, the cover letter should highlight what you find particularly interesting in this topic, why current solutions are limited, as well as describe your overall vision for this project. Ideally, it should also elaborate on why you are interested in working in academic research.
All applications are welcome, regardless of age, gender, social or ethnic origin, sexual orien-
tation, or disability. For the integration of people with disabilities, we are working on possible
adaptations of the positions to be filled – within the limits of the applicable rules for the safety
of people: do not hesitate to contact us to tell us about your situation.
References
[1] E. Coevoet, A. Escande, and C. Duriez. Optimization-based inverse model of soft robots with contact handling. IEEE Robotics and Automation Letters, 2(3):1413–1419, 2017.
[2] F. Faure, C. Duriez, H. Delingette, J. Allard, B. Gilles, S. Marchesseau, H. Talbot, H. Courtecuisse, G. Bousquet, I. Peterlik, et al. Sofa: a multi-model framework for interactive physical simulation. Soft tissue biomechanical modeling for computer assisted surgery:283–321, 2012.
[3] F. Hartmann, M. Baumgartner, and M. Kaltenbrunner. Becoming sustainable, the new frontier in soft robotics. Advanced Materials, 33(19):2004413, 2021.
[4] C. Lee, M. Kim, Y. J. Kim, N. Hong, S. Ryu, H. J. Kim, and S. Kim. Soft robot review. en.
International Journal of Control, Automation and Systems, 15(1):3–15, Feb. 2017. DOI:
10.1007/s12555-016-0462-3.
[5] C. Lichtenthäler and A. Kirsch. Legibility of Robot Behavior: A Literature Review. en.
[6] E. Matsas and G.-C. Vosniakos. Design of a virtual reality training system for human–
robot collaboration in manufacturing tasks. International Journal on Interactive Design
and Manufacturing (IJIDeM), 11:139–153, 2017.
[7] M. Rubagotti, I. Tusseyeva, S. Baltabayeva, D. Summers, and A. Sandygulova. Per-
ceived safety in physical human–robot interaction—a survey. Robotics and Au-
tonomous Systems, 151:104047, 2022. DOI: https://doi.org/10.1016/j.robot.2022.
104047.
[8] R. C. Van Adrichem and J. Jovanova. Human Acceptance As Part of the Soft Robot
Design. en. In ASME 2021 Conference on Smart Materials, Adaptive Structures and In-
telligent Systems, V001T03A007, Virtual, Online. American Society of Mechanical Engi-
neers, Sept. 2021. DOI: 10.1115/SMASIS2021-68268.
[9] Y. Wang, G. Wang, W. Ge, J. Duan, Z. Chen, and L. Wen. Perceived safety assessment
of interactive motions in human–soft robot interaction. Biomimetics, 9(1):58, 2024.
[10] G.-Z. Yang, J. Bellingham, P. E. Dupont, P. Fischer, L. Floridi, R. Full, N. Jacobstein, V.
Kumar, M. McNutt, R. Merrifield, B. J. Nelson, B. Scassellati, M. Taddeo, R. Taylor, M.
Veloso, Z. L. Wang, and R. Wood. The grand challenges of Science Robotics. en. Sci-
ence Robotics, 3(14):eaar7650, Jan. 2018. DOI: 10.1126/scirobotics.aar7650.