Negative efficacy of fixed gain error reducing shared control for training in virtual environments

Abstract

Virtual reality with haptic feedback provides a safe and versatile practice medium for many manual control tasks. Haptic guidance has been shown to improve performance of manual control tasks in virtual environments; however, the efficacy of haptic guidance for training in virtual environments has not been studied conclusively. This article presents experimental results that show negative efficacy of haptic guidance during training in virtual environments. The haptic guidance in this study is a fixed-gain error-reducing shared controller, with the control effort overlaid on the dynamics of the manual control task during training. Performance of the target-hitting manual control task in the absence of guidance is compared for three training protocols. One protocol contained no haptic guidance and represented virtual practice. Two protocols utilized haptic guidance, varying the duration of exposure to guidance during the training sessions. Exposure to the fixed-gain error-reducing shared controller had a detrimental effect on performance of the target-hitting task at the conclusion of a month-long training protocol, regardless of duration of exposure. While the shared controller was designed with knowledge of the task and an intuitive sense of the motions required to achieve good performance, the results indicate that the acquisition of motor skill is a complex phenomenon that is not aided with haptic guidance during training as implemented in this experiment.

The Short-Term Effects of Real-Time Virtual Reality Feedback on Motor Learning in Dance

“…If progress is not made in this area, however, there is a real danger that real-time VR feedback would become counterproductive if uninformed physical trainers use it to direct learners' attention to task-irrelevant features at inappropriate stages in learning (cf. Li et al, 2009;.…”

“…Tabata, Ueda, Kiyofuji, & Kuno, 2005;Drobny, Weiss, & Borchers, 2009). However, caution should be taken when using intuitively designed VR training environments, as it is clear that they may not always benefit motor learning, or reliably facilitate the transfer of motor skills from VR to more naturalistic settings (Li, Patoglu, & O'Malley, 2009). Addressing this issue, researchers have manipulated various practice components, such as virtual fixtures (Rosenberg, 1993), shared control and other haptic feedback mechanisms (Tzafestas, Birbas, Koumpouros, & Christopoulos, 2008), such as vibromotors attached to limbs that cue action within prespecified time constraints (Drobny & Borchers, 2010).…”

Passive and Active Discrimination of Natural Frequency of Virtual Dynamic Systems

“…The findings indicate that haptic feedback, combined with visual feedback, can improve discrimination thresholds for low-frequency dynamic systems perceived passively, while discrimination of the dynamic behavior of actively excited systems does not show sensitivity to feedback modality. This finding will influence shared controller design depending on the approach selected (e.g., record and replay methods that require passive excitation [32], [33], [34] versus virtual fixtures or active shared controllers that require active excitation [4], [5], [7], [8], [31], [35], [36]). The method of excitation alone does not show a significant effect on discrimination ability.…”

“…However, the observed after-effects washed out after a short period of time. Error amplification strategies have been used successfully to speed up human motor learning of a dynamic task [6], [7], though error reduction in a dynamic target-hitting task showed negative training efficacy [8].…”

Preparing drivers for dangerous situations: A critical reflection on continuous shared control

“…Several studies have demonstrated the guidance hypothesis for shared control and other forms of haptic guidance [54]- [58]. A classical example of the guidance hypothesis for haptic guidance (in this case non-overridable physical feedback) is illustrated in Figure 2.…”