Research Papers

Exploring the Dimensions of Haptic Feedback Support in Manual Control

[+] Author and Article Information
D. A. Abbink1

BioMechanical Engineering, Delft University of Technology, Mekelweg 2, 2628 CD Delft, The Netherlandsd.a.abbink@tudelft.nl

M. Mulder

Control and Simulation Division, Faculty of Aerospace Engineering, Delft University of Technology, 2600 GB Delft, The Netherlandsmark.mulder@tudelft.nl


Corresponding author.

J. Comput. Inf. Sci. Eng 9(1), 011006 (Mar 03, 2009) (9 pages) doi:10.1115/1.3072902 History: Received October 01, 2007; Revised November 29, 2008; Published March 03, 2009

A promising way to support operators in a manual control task is to provide them with guiding feedback forces on the control device (e.g., the steering wheel). These additional forces can suggest a safe course of action, which operators can follow or over-rule. This paper explores the idea that the feedback forces can be designed not only to depend on a calculated error (i.e., force feedback) but also on the control device position (i.e., stiffness feedback). First, the fundamental properties of force and stiffness feedback are explained, and important parameters for designing beneficial haptic feedback are discussed. Then, in an experiment, the unassisted control of a second-order system (perturbed by a multisine disturbance) is compared with the same control task supported by four haptic feedback systems: weak and strong force feedback, both with and without additional stiffness feedback. Time and frequency-domain analyses are used to understand the changes in human control behavior. The experimental results indicate that—when well designed—stiffness feedback may raise error-rejection performance with the same level of control activity as during unassisted control. The findings may aid in the design of haptic feedback systems for automotive and aerospace applications, where human attention is still required in a visually overloaded environment.

Copyright © 2009 by American Society of Mechanical Engineers
Your Session has timed out. Please sign back in to continue.



Grahic Jump Location
Figure 1

Haptic feedback control scheme. Through physical interaction with the steering wheel the human operator controls θc(t), which is the control input for a second-order controlled system. The system is perturbed by d(t) resulting in error state e(t) that the operator needs to minimize. The operator can have visual feedback from the error states or haptic feedback from the haptic guidance system. All symbols are discussed in the text.

Grahic Jump Location
Figure 2

Force feedback Ff leads to a translation of the passive force-position characteristics of the control device. Ff>0 leads to an upward translation and a shift of the neutral point θ0 to the left. Ff<0 leads to a downward translation and a shift of the neutral point to the right.

Grahic Jump Location
Figure 3

Stiffness feedback K leads to a rotation of the passive force-position characteristic around the zero-force axis. In the presence of an offset force F0≠0, the rotation moves the neutral point θ0 back toward the origin.

Grahic Jump Location
Figure 4

The combination of force (Ff) and stiffness feedback (K) requires an additional force Fs to ensure the neutral point θ0 remains in the location resulting from the initial shift caused by Ff.

Grahic Jump Location
Figure 5

Frequency response function (gain only) of the error-rejection performance of the force feedback controller (cs=0) and force-stiffness feedback controller (cs=8). The response of the haptic feedback systems was measured without human intervention, for two perturbation bandwidths (solid line <0.5 Hz, dashed line <1.0 Hz). The results are shown for four force gains (cf={0.005;0.01;0.02;0.03}) and for two stiffness gains (cs={0;8}).

Grahic Jump Location
Figure 6

Fixed-base driving simulator setup that was used for the experiment

Grahic Jump Location
Figure 7

Illustration of the experiment display (top), steering wheel (middle), and corresponding feedback situation (bottom). The left side of (a) shows an error requiring a clockwise movement of the steering wheel. The right side of (b) shows an error requiring a counterclockwise movement of the steering wheel.

Grahic Jump Location
Figure 8

Time-domain results over all subjects of three metrics: control activity (top), control effort (middle), and performance (bottom). The feedback conditions are: no feedback (No), force feedback with a low gain (F1), force feedback with a high gain (F2), and both force feedbacks with additional stiffness feedback (KF1, KF2). The results for no visual feedback are shown in the left pane (a) and (c) and for visual feedback in the right pane (b) and (d). The results for different disturbance bandwidths are shown in the top (a) and (b) for D=0.5 Hz and for D=1.0 Hz in the bottom two graphs (c) and (d).

Grahic Jump Location
Figure 9

Frequency response function of a typical subject’s disturbance-rejection performance, for perturbations with low bandwidth (<0.5 Hz, dashed) and the high bandwidth (<1.0 Hz, solid). For the sake of brevity, only the gain is shown. The leftmost plot shows the baseline condition: normal control without haptic feedback. The next two plots show the results (for low and high force gain cf, respectively) when subjects were controlling the system with both visual feedback on with haptic feedback. The last two panels show the same for when subjects controlled without visual feedback. The last four panels show the results for pure force feedback (F, black line), and for force-stiffness feedback (FK, gray line).




Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In