Devices and methods for tactile/haptic output

Many devices have been developed in the last 30 years in order to address the somatic senses of the human operator, but only few have become widely available. The most probable reason for that is that the devices are either not very useful or really expensive (from US $10,000 up to more than US $1,000,000). By ``devices for tactile/haptic output'' we mean devices that have been especially designed for this purpose. In some sense, a standard keyboard and mouse do also provide some kind of haptic feedback, namely the socalled breakaway force when a key or button, respectively, has been pressed. Although this is important as tests have proven to increase the input rate in the case of a keyboard, we do not consider these devices within this section.

Devices with tactile, haptic, or force output address the somatic senses of the user (see 2.1.4 ). This can be done by the following methods (taken from [309]):

Pneumatic stimulation

This can be achieved by air jets, air pockets, or air rings. Problems arise due to muscular fatigue and the pressure or squeezing effect which means that the ability to sense is temporarily disabled. Another drawback of pneumatic devices is its low bandwidth.

Vibrotactile stimulation

Vibrations can either be generated by blunt pins, voice coils, or piezoelectric crystals. These devices seem to be the best ones to address somatic senses because they can be build very small and lightweight and can achieve a high bandwidth.

Electrotactile stimulation

Small electrodes are attached to the user's fingers and provide electrical pulses. First results are promising, but further investigation is needed in this area.

Functional neuromuscular stimulation (FMS)

  In this approach, the stimulation is provided directly to the neuromuscular system of the operator. Although very interesting, this method is definitely not appropriate for the standard user.

Nearly all devices with tactile output have been either developed for graphical or robotic applications. Many different design principles have been investigated, but the optimal solution has not been found yet. Most probably, the increasing number and popularity of Virtual Reality systems will push the development of force feedback devices to a new dimension. In the following, the most popular devices will be reviewed very briefly in chronological order:

The Ultimate Display (1965)

Ivan Sutherland described his vision of an ``ultimate display'' in order to reflect the internal world of the machine as close as possible. He proposed to develop a force reflecting joystick [334].

GROPE (1967 -- 1988)

In the project GROPE, several ``haptic displays'' for scientific visualization have been developed in different stages of the project [48]. Starting in 1967, a 2D device for continuous force feedback (similar to an X/Y-plotter) has been developed. In the next stage (1976), a 6D device was used in combination with stereo glasses for operations in 3D space. The device was a kind of master manipulator with force feedback. In the third (and last) phase of GROPE which started in the late 1980s, the hardware has been improved, but the principle layout of the system was not changed. The results in the domain of molecular engineering seem to be very promising, i.e. the performance has been increased significantly by the use of haptic feedback.

Joystring (1986)

  A very special design of a 6D manipulator with force feedback has been realized independently by Agronin [4] and Staudhamer (mentioned in [103], further developed by Feldman). A T-shaped grip is installed inside a box with three strings on each of its three ends. In Agronin's version, the strings' tension is controlled by three linear motors, whereas the second one uses servo motors. No results have been published.

The Exoskeleton (1988)

For telerobotic applications, different skeletons which have to be mounted on the operators arm have been built at the JPL [150], the EXOS company [87], and the University of Utah (in [269]). The systems are used as a kind of master-slave combination, and forces are applied by motors at the joints. Unfortunately, these devices are usually very heavy, therefore they can also be used in special applications. EXOS, Inc. has developed a ``light'' version for the NASA, but this system does not have any force feedback.

The Compact Master Manipulator (1990)

The master manipulator that is presented in [148] is based on flight simulator technique. All three translational and rotational axis can be controlled and are equipped with force feedback. Additionally, the thumb, the indexfinger, and the other three fingers control a grip which also applies forces to the user, thus yielding a 9-DOF device with force feedback. Unfortunately, the paper only covers the design of the manipulator but does not contain any results.

A 3D Joystick with Force Feedback (1991)

A joystick for positioning tasks in 3D space has been developed by Lauffs [174]. The 3 axis of the stick can be controlled independently, but rotations are not possible. Force feedback has been realized by the use of a pneumatic system, which is very robust but too slow for most applications.

A Force Feedback Device for 2D Positioning Tasks (1991)

A device that is very similar to the one developed in GROPE I (see above) has been realized by Fukui and Shimojo [111]. Instead of a knob, the X/Y-recorder is moved with the finger tip. The resulting force will be calculated and sent to the application, and if a collision is detected, one or both axis will be blocked. This device has been developed for contour tracking operations, its advantage is its almost friction and mass free operation.

PUSH (1991)

In [131], a Pneumatic Universal Servo Handcontroller is described. It has been developed for the control of industrial robots. By using cardan joints and an orthogonal coordinate system which is placed in the operator's hand, all axis can be controlled independently. The device, which is rather large, can apply forces by pneumatic cylinders. As the 3D joystick described above, it is very slow.

Teletact (1991)

  A data glove with tactile feedback has been developed by Stone [326] and is used for outputs to the user, whereas a second data glove is used for inputs to the computer. The input glove is equipped with 20 pressure sensors, and the output glove with 20 air pads, controlled by 20 pneumatic pumps. The major drawback of this system is the very low resolution, additionally a force feedback is missing completely. The next generation, Teletact-II, has been equipped with 30 air pads and is available on the market now.

Force Dials (1992)

A dial with force feedback has been realized by computer scientists and chemists for simple molecular modeling tasks [127]. The force is controlled by a motor. The main advantage of this device is its low price and its robustness but due to its simplicity it will not be very useful for a multimodal system.

Multimodal Mouse with Tactile and Force Feedback (1993)

In [6], an interesting approach to equip a standard input device with output capabilities has been described. A common mouse has been equipped with an electro magnet and a small pin in its left button. This idea is especially appealing because it is cheap and easy to realize. First results are very promising and have shown that the performance in positioning tasks can be increased with this kind of feedback by about 10%.

PHANToM (1993)

In his master thesis, Massie developed a 3D input device which can be operated by the finger tip [214]. It realizes only translational axis, but it has many advantages compared to other devices, like low friction, low mass, and minimized unbalanced weight. Therefore, even stiffness and textures can be experienced.

A 2D Precision Joystick with Force Feedback (1994)

Sutherland's basic idea has been realized by researchers at the University of New Brunswick [10]. The joystick has been made very small in order to achieve a very high precision in its control. Results have shown that the accuracy in a contour modification task can be increased (44%), but the time will increase (64%), too.

Esprit Project 8579/MIAMI (Schomaker et al., '95)
Thu May 18 16:00:17 MET DST 1995