PAPER

DESIGN AND TEST OF A 1 DOF HAPTIC DEVICE FOR ONLINE EXPERIMENTATION

Design and Test of a 1 DOF Haptic Device for

Online Experimentation
http://dx.doi.org/10.3991/ijoe.v12i04.5152

F. Carneiro, M. R. Quintas, P. Abreu and M. T. Restivo

University of Porto, Porto, Portugal

Abstract—This work describes the design and testing of a

1 degree of freedom (DOF) haptic device, for interaction II. SYSTEM DESIGN
with virtual reality (VR) applications. This device was de- We propose a device that allows linear movement of a
signed so that it could to be easily assembled by anyone at a slider knob along one direction, capable of applying ad-
reduced price. Instructions for device assembly, as well as justable bi-directional force in any point of its working
some VR applications are available online. The tests per- range [6]. It is operated by manually grasping and moving
formed on the device proved that, despite its low cost, it is the knob, through which the force sensation is transmitted.
able to provide users with a good haptic experience and is This design was based on a previous project, in which a
adequate for not complex applications. similar device was built from components of obsolete
equipment [7].
Index Terms—1 DOF haptic device, haptic interaction,
virtual reality.
The device was designed under the principle that it
should be easy to reproduce at a low cost. Thus, the use of
low-cost and off-the-shelf components, easily found on
I. INTRODUCTION
the internet, was another self-imposed guideline. Moreo-
One notable means of increasing immersion in virtual ver, by resorting to a 3D printer, it is possible to design
reality environments is by making use of sensorial devices functional parts with complex shapes that can be produced
such as haptic interfaces [1, 2]. at a low cost. The system should also be easy to disassem-
In a broader sense, the term haptic relates to the study ble, to allow replacement of worn out or malfunctioning
of touch and tactile sensations [3]. As such, a haptic inter- parts.
face can be defined as a system capable of providing tac- A small DC motor drives the slider knob: a toothed pul-
tile and/or force feedback. These interfaces typically con- ley, connected to the motor shaft, engages a timing belt
sist of electromechanical devices and have applications in whose ends are joined to a carriage, on which the knob is
numerous fields, including medicine, research, training, attached. The carriage includes linear bearings and slides
education, industry and aeronautics, etc. [4]. There are on two metal guiding shafts. This combination provides
currently several haptic devices on the market, mostly the device with great capacity to withstand disturbing
with three or six degrees of freedom (DOF). These are forces, while minimizing friction and ensuring high dura-
available at a cost that can prove prohibitive for institu- bility. To keep the belt under tension we used an idler
tions or individuals intending to perform not complex wheel, comprised of two flanged ball bearings, mounted
haptic simulations, which could be performed with sim- on another metal shaft. Belt tension is adjusted through a
pler devices. screw mechanism on the carriage. The knob also contains
The Hapkit is one example of a low-cost 1 DOF haptic two ball bearings, to allow its rotation without increasing
device. This device, developed by a team at Stanford Uni- friction on the guides. An encoder mounted on the motor
versity, was designed to be easily assembled [5]. It inter- shaft allows indirect measuring of the carriage position. A
acts with the user though a 1 DOF joystick-like handle, depiction of the transmission mechanism is presented in
that performs rotational movement. This kind of move- Fig. Error! Reference source not found..
ment, as well as the low output force it produces, limits The motor and pulley were chosen to provide the de-
the range of applications for the device. To the best of our vice with a force range of 20 N. The linear working range
knowledge, there are no commercially available haptic was fixed as 100 mm for knob movement. The overall
devices with one linear DOF, although some suppliers dimensions for the device were set to be around
admit to build such systems on demand, at a higher price. 200!100!50 mm.
Therefore, we identified the need to develop a 1 DOF The electronics for the control circuit in the device have
haptic device capable of providing linear movement with also been further optimized, in order to lower their total
high output force.
We designed this solution in order to be available as an
inexpensive and easy to assemble do-it-yourself (DIY) kit,
with the intention of spurring people to become more
acquainted with haptic devices.
Furthermore, we tested the device to establish some of
its characteristics, so that it could be compared with other
haptic interfaces.
Figure 1. Side cross-section of the device, modeled in CAD software

iJOE ‒ Volume 12, Issue 4, 2016 55

 PAPER
DESIGN AND TEST OF A 1 DOF HAPTIC DEVICE FOR ONLINE EXPERIMENTATION

cost, and achieve greater performance and efficiency in a

smaller footprint.
A computer, that interfaces through a USB connection
reads the knob position and sends the matching force
value to the device, depending on the running software.
Finally, we estimate that the total cost of materials for
the device is under 80".

III. ONLINE SERVICES

Some software applications that can be used with the
haptic device are available at [8], where they are referred
to as “VR applications with haptic interaction”. On the
same webpage all the files and information necessary for
replicating the device are available, under “do it yourself 1
DOF haptic” (Fig. Error! Reference source not found.).
These include:
• Every file for 3D printing the device (in STL
format);
• A complete part list, detailing the required in-
a)
formation for procuring the remaining parts;
• An assembly manual, instructing how to as-
semble the device;
• An assembly video, generated resorting to
CAD software;
• A full schematic and component list for the
electronics;
• A video, displaying how the haptic device in-
teracts with one of the available applications.
The electronic control circuit can also be assembled by
anyone; however, we must supply them with the prepro-
grammed microcontrollers.

IV. TESTING AND RESULTS

From a functional standpoint, the main guideline fol- b)
lowed when designing the new version of the 1 DOF hap- Figure 2. Example of the information provided on the website: (a)
tic device, was that it should perform as well or better than available files and (b) the assembly video
the previous prototype. Throughout development, this
parameter was regularly assessed by testing each iteration
of the device with the applications it was supposed to
interface with. Upon reaching the final prototype (Fig. 3
and realizing we were satisfied with the interactions it
allowed, we decided to further characterize the device, in
order to fully understand where its capabilities and limita-
tions lie.
In spite of the increasing popularity of haptic device
development [4], there are still no established metrics for
universal performance comparison across haptic devices
[9-11]. As such, parameters for testing were chosen from
[9] and [11], based on their effect on actual usage of the
device and ease of implementation. The tested parameters
were:
• Static friction force; Figure 3. Picture of the final device prototype
• Peak force:
o Short transient peak force; A. Testing methodologies
o Long term peak force; Static friction was measured with a load cell that was
• Peak acceleration; placed between the hand and the handle of the device.
With the motor unpowered, the load cell was progressive-
• Total inertia;
ly pushed until the encoder registered movement, keeping
• Peak velocity. record of the maximum exerted force.

56 http://www.i-joe.org
 PAPER
DESIGN AND TEST OF A 1 DOF HAPTIC DEVICE FOR ONLINE EXPERIMENTATION

For peak force measurement, the device was placed on REFERENCES

its side, so that the knob moved vertically. The motor was [1] M. G. Jones, A. Bokinsky, T. Andre, D. Kubasko, A. Negishi, R.
set to maximum voltage (in order to exert maximum Taylor, et al., "Nanomanipulator applications in education: the
force) and the knob was loaded with different known impact of haptic experiences on students' attitudes and concepts,"
masses. For each mass value we checked if handle move- in Haptic Interfaces for Virtual Environment and Teleoperator
ment occurred. In the case of short transient peak force, Systems, 2002. HAPTICS 2002. Proceedings. 10th Symposium on,
2002, pp. 279-282. http://dx.doi.org/10.1109/haptic.2002.998971
the motor was allowed to cool down to room temperature
[2] C. Basdogan, C.-H. Ho, M. A. Srinivasan, and M. Slater, "An
between each trial, while for long term peak force the
experimental study on the role of touch in shared virtual environ-
device was tested continuously. ments," ACM Trans. Comput.-Hum. Interact., vol. 7, pp. 443-460,
We measured peak acceleration with a MEMS accel- 2000. http://dx.doi.org/10.1145/365058.365082
erometer (MPU-6050) attached to the handle of the de- [3] T. Sheridan, "Human and machine haptics in historical perspec-
vice, which was set to run at maximum voltage. A micro- tive," in Workshop on Human and Machine Haptics, 1997.
controller acquired the acceleration data and transmitted it [4] V. Hayward, O. R. Astley, M. Cruz!Hernandez, D. Grant, and G.
to a computer through a USB connection, where it was Robles!De!La!Torre, "Haptic interfaces and devices," Sensor Re-
view, vol. 24, pp. 16-29, 2004. http://dx.doi.org/10.1108/026022
displayed using custom software. 80410515770
Total inertia or apparent mass at tip (knob) was calcu- [5] Stanford University. (2014). Hapkit. Available:
lated from the peak force and peak acceleration values, http://hapkit.stanford.edu/
through (1). [6] F. Carneiro, J. Silva, M. Quintas, P. Abreu, and M. T. Restivo, "1
DOF Haptic Device for Online Experimentation," in 3rd Experi-
!"#$!!"#$% ment@ International Conference, University of Azores, Ponta
!"#$%&' ! ! (1) Delgada, Portugal, 2015, pp. 143-144.
!"#$!!""#$#%&'()* [7] P. U. Gomes, M. Rodrigues Quintas, and M. T. Restivo, "1 DOF
haptic device built with parts of recycled material," in Experi-
Lastly, we measured peak velocity using the encoder ment@ International Conference (exp.at'13), 2013 2nd, 2013, pp.
available on the device. The microcontrollers were pro- 152-153. http://dx.doi.org/10.1109/expat.2013.6703049
grammed so that the encoder pulses were counted every [8] (2015). ONLINE EXPERIMENTS LIM. Available:
0.002 seconds and the maximum value was sent to a com- http://onlinelab.fe.up.pt/vr_exp.html
puter via USB, where it was displayed using custom soft- [9] V. Hayward and O. Astley, "Performance Measures for Haptic
ware. Interfaces," in Robotics Research, G. Giralt and G. Hirzinger,
Eds., ed: Springer London, 1996, pp. 195-206.
B. Test Results and Discussion [10] E. Samur, W. Fei, U. Spaelter, and H. Bleuler, "Generic and
A summary of the results obtained from the performed systematic evaluation of haptic interfaces based on testbeds," in
Intelligent Robots and Systems, 2007. IROS 2007. IEEE/RSJ In-
tests is displayed on Table 1. ternational Conference on, 2007, pp. 2113-2119.
The obtained results are within the expected range and [11] R. E. Ellis, O. M. Ismaeil, and M. G. Lipsett, "Design and evalua-
confirm the quality of the designed haptic device. tion of a high-performance haptic interface," Robotica, vol. 14, pp.
The device exhibits good backdriveability, with low in- 321-327, 1996. http://dx.doi.org/10.1017/S0263574700019639
ertia and static friction comparable to other high perfor- [12] M. Moreyra and B. Hannaford, "A practical measure of dynamic
response of haptic devices," in Robotics and Automation, 1998.
mance devices [11, 12]. The low static friction to peak Proceedings. 1998 IEEE International Conference on, 1998, pp.
force ratio is also extremely positive, since it assures the 369-374 vol.1. http://dx.doi.org/10.1109/robot.1998.676430
device has a good dynamic range. The remaining results [13] R. Q. Van der Linde, P. Lammertse, E. Frederiksen, and B. Ruiter,
are also within range of what is expected from a haptic "The HapticMaster, a new high-performance haptic interface," in
interface [11, 13, 14] and reassure us that the device Eurohaptics, 2002.
should provide good fidelity for haptic interactions. [14] L. Stocco and S. E. Salcudean, "A coarse-fine approach to force-
reflecting hand controller design," in Robotics and Automation,
V. CONCLUSIONS 1996. Proceedings., 1996 IEEE International Conference on,
1996, pp. 404-410 vol.1. http://dx.doi.org/10.1109/robot.
In this paper, a 1 DOF haptic device that can be assem- 1996.503810
bled by anyone at a low cost was proposed. This device
was also tested, using a set of defined parameters. The AUTHORS
tests revealed the device is capable of providing good F. Carneiro is with INEGI research center, University
haptic interactions, which is remarkable, especially con- of Porto, Porto, Portugal ([email protected]).
sidering its low cost and ease of assembly.
M. R. Quintas, P. Abreu and M. T. Restivo are with
This device has been tested with children at the K-12 the UISPA-LAETA/INEGI research center and with Fac-
level, with promising results. ulty of Engineering of University of Porto, Porto, Portugal
Instructions and other materials for replicating the de- ([email protected], [email protected], [email protected]).
vice are available on-line, which we hope might encour-
This work was funded by the National Science Foundation (FCT –
age people to replicate it and further promote the use of Portuguese Foundation for Science and Technology) under the project
haptic interfaces. UID/EMS/50022/2013. Submitted 19 October 2015. Published as resub-
mitted by the authors 13 December 2015.

iJOE ‒ Volume 12, Issue 4, 2016 57