Haptic Wrists: An Alternative Design Strategy Based on User Perception Javier Martín1, Joan Savall2, Iñaki Díaz2, Josune Hernantes2 & Diego Borro2 1

Centro de Estudios e Investigaciones Técnicas (CEIT), Department of Applied Mechanics, Paseo Manuel Lardizábal 15, 20018, San Sebastian, Spain 2 Centro de Estudios e Investigaciones Técnicas (CEIT), Department of Applied Mechanics, and Tecnun (University of Navarra), Paseo Manuel Lardizábal 13, 20018, San Sebastian, Spain

Abstract A new 3 degrees-of-freedom (DOF) torque feedback wrist is being developed to be added to an existing 3 DOF force feedback haptic device. It is difficult to find a satisfactory solution to the mechanical design problem, mainly because of the required large rotational workspace and severe weight constraints. This work proposes an alternative design strategy based on user perception which allows simplification of the mechanics. The proposed approach consists of substituting the last rotational DOF of the wrist with a pseudo-haptic DOF. Thanks to specially designed visuotactile cues, the pseudo-haptic DOF is integrated with the active DOF into the same device, being able to generate free motion and collision detection perception to the user. This approach provides for simpler kinematics, lightweight designs, lower inertias and less friction, which are key advantages for the inclusion of torque feedback into force feedback devices.

1

Introduction and motivation

Haptic interfaces are both i) input devices which capture the movements of a user and ii) feedback devices which allow the user to physically interact with a virtual or remote environment. Some haptic applications require a large-scale workspace. One example is the Large Haptic Interface for Aeronautics Maintainability (LHIfAM) developed at CEIT [1, 2]. LHIfAM measures 6 DOF translation and orientation and provides 3 DOF force feedback, allowing the user to interact with full-scale 3D virtual aircraft engines for maintainability analysis purposes. Currently LHIfAM is being augmented to 6 active DOF in order to provide both force and torque feedback. This demands a lightweight torque feedback wrist providing high output torques within a large rotational workspace. A state-of-the-art in mechanisms for haptic torque feedback was presented by the authors [3], concluding that none of the reviewed wrists suits our needs. An in-depth study of the challenges presented by torque feedback was also carried out by the authors [4] showing that, design-wise, there is no way to simultaneously guarantee a large rotational workspace with high quality performance. In order to avoid the severe mechanical limitations, this work moves the scope of the design solutions closer to the user’s perception through pseudo-haptic and sensory substitution techniques. This allows simplification of the mechanics. The paper is organized as follows: Section 2 summarizes previous work on pseudo-haptics and sensory substitution. The concept, implementation, applicability and limitations of the proposed approach are explained in Section 3. Finally, conclusions and future work are presented in Section 4.

2

Related work

The first choice for simplifying the 6 DOF mechanics is to simulate torque with the pure force mechanism itself, without adding any wrists. In that respect, Lécuyer et al. [5] propose an interaction paradigm called the “Virtual Haptic Sphere” which allows one to use the PHANToMTM alternately as a 3 DOF input/output device in translation and as a 2 DOF input/output device in rotation. Another work [6] argues that even without true torque feedback exerted by the haptic device, the user can voluntarily supply the sense of torque, aided by visualization and previous experience with the task being performed. This is emphasized by [7] with further experiments displaying haptic information with both torque and visual feedback. His work poses the following question: “Can the

same level of haptic perception be maintained if the number of degrees of freedom of the haptic interface is reduced?” The impact of visually presented spatial cues on the human perception of mechanical stiffness in virtual environments is studied in [8], demonstrating a clear visual dominance over the kinesthetic sense of hand position. Based on this idea and previous work [9] with isometric and elastic/isotonic input devices1 [10] for target acquisition tasks in virtual environments, Lécuyer et al. [11] propose the “pseudo-haptic” concept. This preliminary state consists of coupling a SpaceballTM input device with a visual feedback, displaying a visual stiffness different from the mechanical stiffness of the device. This results in the user being able to perceive different stiffness values by means of a misleading visual stiffness display, even if the mechanical stiffness is always the same. Thus, a passive input device like a SpaceballTM can, to a certain extent, simulate force feedback. Going more in depth into this concept with exhaustive experimentation, Lécuyer et al. [12] show the unconscious participation of the user in this phenomenon. It is linked, on the one hand, to the user’s cognitive strategy during the experimental task and, on the other hand, to the role of sensory motor commands in the user’s own perception loop. Later, the same authors extend the pseudo-haptic definition to “the generation, augmentation or deformation of haptic sensations by information coming from other sensory modalities” [13], concluding that integration of pseudo-haptic and haptic feedback in the design of VEs is possible. All of these concepts come together on a virtual milling machine technical trainer [14]. Regarding torque feedback, Paljic et al. [15] validate the applicability of the pseudo-haptic concept to 1 DOF torque feedback for torsional stiffness discrimination, stating that elastic devices provide a better resolution than isometric ones. Closely related to these phenomena, “sensory substitution” is defined by [16], and later studied by [17], as “the provision to the brain of information that is usually in one sensory domain (for example visual information via the eyes and visual system) by means of the receptors, pathways and brain projection, integrative and interpretative areas of another sensory system, (for example visual information through the skin and somatosensory system)”. In virtual assembly tasks, [18] investigated whether the substitution of force feedback with auditory cues improved manipulation performance and subjective perception of usability. They found that, depending on the specific goal (minimize number of collisions, increase user perception, etc.), this sensory substitution could be valuable. They also described a number of guidelines for designers to that effect. In the medical field, [19] analyzed the effects of substituting direct haptic feedback with visual and auditory cues to provide surgeons with a representation of force magnitude in teleoperated suture-manipulation procedures. They showed that sensory substitution is capable of providing sufficient feedback for the user to control these robotically-applied forces. The potential of the sensory substitution for force feedback through tactile and auditory senses was studied by [20], concluding that “sensory substitution can be considered as a proven method by which force information can be presented if traditional force feedback is too costly, impractical, inefficient, or unstable for the given task conditions”. In particular, his experiments showed that sensory substitution through tactile and auditory senses was well suited for tasks that were highly dependent on reaction time, like detecting the presence of contact forces. Taking all these concepts into account, we believe i) design advantages can be achieved by integrating haptics and pseudo-haptics into a single feedback device, and ii) design guidelines derived from studies on sensory substitution can be helpful in making such integration more effective.

3

”2.5 DOF” haptic torque feedback approach

Ergonomically, tool manipulation can be divided into a pointing movement (pitch/yaw) and a rotation along the tool axis (tool-roll). Pointing movement of the tool involves certain arm-wrist movements which are different from those needed for tool-roll rotation. As a matter of fact, it can be observed that, in haptics, dealing with tool-roll separately from the pitch/yaw movements is a

1

Isometric devices are designed to display displacement variables and read back forces, so they have ideally infinite stiffness and stay put while force is exerted on them. Elastic/Isotonic devices are designed to display forces and read position variables, thus they offer no significant resistance and are used to track users as they move within the virtual world (from [10], [11]).

common mechanical design philosophy [3]. The alternative wrist design proposed in this paper consists of the substitution of the tool-roll haptic DOF with a tool-roll pseudo-haptic DOF (Figure 1).

Figure 1: Proposed 2.5 DOF approach.

This mixture of 2 haptic DOF plus 1 pseudo-haptic DOF is what the authors refer to as the 2.5 DOF2 haptic torque feedback approach: a wrist with 2 active DOF (pitch/yaw) plus an additional sensor which measures the torque exerted by the user along the handle axis, which coincides with the tool-roll axis. By design, there is no physical rotation of the handle along the tool-roll axis, and measured torque is instead used for interaction with the virtual environment (isometric DOF). In order to give the impression of 3 active DOF with only 2 actuated DOF, “it is necessary to facilitate the repositioning of the user perception to an implicit solution. It means that this implicit sensory alternative must be explicit enough to be found quickly by the user” [12]. Following this idea and inspired by the concepts reviewed in Section 2, a combination of visual and tactile cues (henceforth “visuotactile cues”) has been designed to convey user perception of tool-roll DOF. The designed visuotactile cues follow the criteria stated by Massie et al. [22] for a haptic device to be an effective interface: i) free space must feel free and ii) solid virtual objects must feel stiff. First, as rotation along the tool-roll axis is physically impossible, torque (T) is involuntarily exerted along the handle axis when the user tries to make such movement. This is measured by the torque sensor and then translated into a visual rotation of the virtual tool (with rotation speed ω), displayed to the user through visual feedback (Figure 2).

Figure 2: Tool-roll axis, free motion perception.

As seen in Eq. (1), virtual tool-roll rotation speed (ω) is set proportional (a) to a power (b) of the measured tool-roll torque (T), with sign(ω) equal to sign(T). As b increases, precision manipulation produces slower rotations and power grip produces faster rotations. In addition, rotation speed is limited to a maximum realistic value (ωmax). The parameters of the function need to be tuned to achieve a free space perception as similar as possible between the real DOF and the pseudo-haptic DOF. This tuning is particular for each mechanical device and application.

ω = a T sign (T ) , ω ≤ ωmax b

(1)

When collision with a solid virtual object occurs, a tactile cue is sent to the user. In addition, as a visual cue, interpenetration between virtual tool and environment is avoided in the visual feedback. The visual cue suitably drives the tactile cue to convey collision perception to the user. The tactile cue 2 The term “2D and half” has been used before [21] in a slightly different sense: the PenCat/ProTM device is a planar five bar linkage which allows a measured vertical movement of the pen, passively actuated by elastic return. “2.5 DOF” is used in this paper for simplicity.

consists of a single pulse of short duration (∆) and small magnitude (M), given off at the moment of collision (tc), and it is exerted by the pitch/yaw actuators (Figure 3). Thus, collisions due to pitch/yaw rotation produce kinesthetic torque feedback and collisions due to tool-roll rotation produce tactile feedback. Collisions involving all rotational DOF produce both kinds of feedback superimposed. Due to the short duration and small magnitude of the pulse, tactile feedback does not affect pitch/yaw kinesthetic feedback when both are superimposed. Moreover, since all feedback is carried out by the pitch/yaw actuators, no additional actuator is needed to produce the pulse. The magnitude of the pulse is set proportional to the virtual tool-roll rotation speed (ω) at the collision instant, and it can be invariably directed to the same actuator. Again, the parameters of the pulse need to be tuned depending on the mechanical device and the application.

Figure 3: Tool-roll axis, collision perception.

The 2.5 DOF haptic wrist provides for simpler kinematics, lightweight designs, lower inertias and less friction, which are key advantages for the design of torque feedback wrists to be added to force feedback devices. On the other hand, the proposed approach has two main limitations: i) It works best with longitudinal virtual tools and it requires the tool-roll, the device handle and the longitudinal axis of the virtual tool to be aligned, and ii) since the tool-roll is an isometric DOF, it provides collision detection ability but not stiffness discrimination. However, even if tool-roll rotation is not mechanically permitted by the device, we observed that the deformation of the skin at the contact points allows the rest of the hand to undertake a certain tool-roll rotation, which helps creating a realistic sensation of physically turning the hand.

4

Conclusions and future work

A new approach has been proposed for the mechanical design of haptic wrist mechanisms, consisting of the substitution of the wrist tool-roll active DOF with a pseudo-haptic DOF. Especially designed visuotactile cues integrate haptic and pseudo-haptic torque feedback into the same device, providing the impression of 3 active DOF with only 2 actuated DOF. This gives rise to a “2.5 DOF” wrist. Such strategies have been explained, and advantages and disadvantages of the new approach have been pointed out. The 2.5 DOF haptic wrist does not intend to replace a full torque feedback wrist, but its mechanical simplification helps the inclusion of torque feedback into force feedback devices. Experiments are being carried out to validate the applicability of the new approach in VE interactions. So far, we believe this approach introduces an alternative design strategy, spans new research lines and may help increase knowledge in the field of the human haptic torque perception problem.

5 [1]

References Savall, J., Borro, D., Gil, J. J., and Matey, L., 2002, "Description of a Haptic System for Virtual Maintainability in Aeronautics," Proc. International Conference on Intelligent Robots and Systems, IEEE/RSJ, EPFL, Lausanne, Switzerland, 3, pp. 2887-2892.

[2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14]

[15]

[16] [17] [18] [19] [20] [21] [22]

Borro, D., Savall, J., Amundarain, A., Gil, J. J., García-Alonso, A., and Matey, L., 2004, "A Large Haptic Device for Aircraft Engines Maintainability," IEEE Computer Graphics and Applications, 24(6), pp. 70-74. Martin, J., and Savall, J., 2005, "Mechanisms for Haptic Torque Feedback," Proc. WorldHaptics'05, IEEE, Pisa, Italy, pp. 611-614. Gosselin, F., Martín, J., Andriot, C., and Savall, J., 2007, "Large Workspace Haptic Devices for Human Scale Interaction," Proc. 1st International Workshop on MultiModal Interfaces for the Transfer of Human Skills, SKILLS IP (IST-2006-035005), Pisa, Italy. Lécuyer, A., Kheddar, A., Coquillart, S., Graux, L., and Coiffet, P., 2001, "A Haptic Prototype for the Simulations of Aeronautics Mounting/Unmounting Operations," Proc. International Workshop on Robot-Human Interactive Communication, IEEE, Bordeaux and Paris, France. Hancock, D., 1996, "The Sense of Torque with a Single Phantom Haptics Device," Proc. 1st PHANToM User's Group Workshop, Dedham, MA, MIT AI Lab TR 1596. Wang, S., 2001, “Role of torque in haptic perception of virtual objects,” Thesis, Massachusetts Institute of Technology, Mechanical Engineering Dept.. Srinivasan, M. A., Beauregard, G. L., and Brock, D. L., 1996, "The Impact of Visual Information on the Haptic Perception of Stiffness in Virtual Environments," Proc. Dynamics Systems and Control Division, ASME, 58, pp. 555-559. Zhai, S., 1993, "Investigation of Feel for 6dof Inputs: Isometric and Elastic Rate Control for Manipulation in 3D Environments," Proc. The Human Factors and Ergonomics Society 37th Annual Meeting. Hayward, V., and Astley, O. R., 1996, "Performance Measures for Haptic Interfaces," Proc. Robotics Research: The 7th Int. Symposium, G. Giralt and G. Hirzinger, eds., Springer Verlag, pp. 195-207. Lécuyer, A., Coquillart, S., Kheddar, A., Richard, P., and Coiffet, P., 2000, "Pseudo-Haptic Feedback: Can Isometric Input Devices Simulate Force Feedback?" Proc. International Conference on Virtual Reality, IEEE, pp. 83-90. Lécuyer, A., Coquillart, S., and Coiffet, P., 2000, "Simulating Haptic Information with Haptic Illusions in Virtual Environments," Proc. NATO RTA/Human Factors & Medicine Panel Workshop, The Hague, The Netherlands. Lécuyer, A., Burkhardt, J. M., Coquillart, S., and Coiffet, P., 2001, "`Boundary of Illusion´: an Experiment of Sensory Integration with a Pseudo-Haptic System," Proc. International Conference on Virtual Reality, IEEE, Yokohama, Japan, pp. 115-122. Crison, F., Lécuyer, A., Mellet-D’Huart, D., Burkhardt, J. M., Michel, G., and Dautin, J. L., 2005, "Virtual Technical Trainer: Learning How to Use Milling Machines with Multi-Sensory Feedback in Virtual Reality," Proc. International Conference on Virtual Reality, IEEE, Bonn, Germany, pp. 139146. Paljic, A., Burkhardt, J.M., and Coquillart, S., 2004, "Evaluation of Pseudo-Haptic Feedback for Simulating Torque: a Comparison Between Isometric and Elastic Input Devices," Proc. 12th International Symposium on Haptic Interfaces for Virtual Environment and Teleoperator Systems, IEEE, Chicago, Illinois, pp. 216-223. Bach-y-Rita, P., Webster, J. G., Tompkins, W. J., and Crabb, T., 1987, "Sensory Substitution for Space Gloves and for Space Robots," Proc. Workshop on Space Telerobotics, NASA/JPL, Pasadena, CA, 2, pp. 51-57. Lenay, C., Gapenne, O., Hanneton, S., Marque, C., and Genouëlle, C., 2003, Sensory substitution, limits and perspectives, Touch for Knowing, John Benjamins Publishers, Amsterdam, The Netherlands, pp. 275-292. Edwards, W., Barfield, W., and Nussbaum, A., 2004, "The Use of Force Feedback and Auditory Cues for Performance of an Assembly Task in an Immersive Virtual Environment," J. Virtual Reality, 7, pp. 112-119. Kitagawa, M., Dokko, D., Okamura, A. M., and Yuh, D. D., 2005, "Effect of Sensory Substitution on Suture-Manipulation Forces for Robotic Surgical Systems," J. Thoracic and Cardiovascular Surgery, 129, pp. 151-158. Massimino, M. J., 1995, "Improved Force Perception Through Sensory Substitution," J. Control Engineering Practice, 3, pp. 215-222. Hayward, V., "Survey of Haptic Interface Research at McGill University," Proc. Workshop in Interactive Multimodal Telepresence Systems, TUM, Munich, Germany, pp. 91-98. Massie, T. H., and Salisbury, J. K., "The PHANTOM Haptic Interface: A Device for Probing Virtual Objects," Proc. Symposium on Haptic Interfaces for Virtual Environment and Teleoperator Systems, ASME, Chicago, IL.