Page 19 - Read Online

Jin et al. Soft Sci 2023;3:8 https://dx.doi.org/10.20517/ss.2022.34 Page 17 of 26

Figure 9. Robotics for human-machine interactions (HMIs): (A) tactile glove resistive sensing array with high spatial resolution to
[34]
recognize human grasping gestures (reproduced with permission . Copyright 2019, Springer Nature); (B) human beings manipulate
the high-resolution tactile sensor switch to control the robotic grasper for needle threading (reproduced with permission [30] . Copyright
2021, The American Association for the Advancement of Science); (C) close-looped control of the remote robotic hand to grasp a
softball, human beings utilize the EMG signal to actuate the robotic hand, and the grasping force signal is reversed by electrical
stimulator (reproduced with permission [54] . Copyright 2022, The American Association for the Advancement of Science); (D) digital-
twin-based virtual shop based on VR devices and AI models (reproduced with permission [185] . Copyright 2021, John Wiley and Sons).

manipulation tasks. Numerous flexible human-machine interfaces have been reported, such as traditional
input devices, namely screen, mouse or keyboard , and wearable tactile systems, especially EMG skin
[176]
patches or intelligent data gloves . Advanced algorithms such as CNN, Linear Discriminant Analysis
[178]
[177]
(LDA), Root Mean Square (RMS), and confusion matrixes can assist in precisely extracting human
instructions. These devices have been applicated to remote controlling of the mobile robot ,
[179]
[180]
[19]
wheelchair , robotic grasping manipulator , or virtual game character . Yan et al. have reported a
[181]
high-resolution magnetic tactile sensor array for the teleoperation of a robotic manipulator in Figure 9B .
[30]
NN algorithms can intelligently interpret human finger touch signals into grasper movement commands,
which can even be used for a sophisticated task of needle threading. However, in most cases, robotics just
functions as a receiving terminal for data processing and command execution without tactile feedback.
Human manipulators can only observe the operation process and give instructions regardless of on-site
perception and cognition, which makes it challenging to make corresponding adjustments according to the
actual operation state of the robotics.