Source: edgy.app
A team of researchers has developed a predictive touch response mechanism for enhanced human-to-machine Interaction.
Experts have predicted that the next phase of IoT could be Tactile Internet. So, humans will be able to interact with a remote or virtual object and experience realistic haptic feedback.
Now a team of researchers led by Elaine Wong at the University of Melbourne has brought us closer to that future.
The researchers invented a method for enhancing feedback experience in human-to-machine applications.
Aside from Tactile Internet, the process can also predict feedback in other applications as well. These include electronic healthcare and virtual reality gaming.
Wong and her colleagues intend to present their proposed module at the Optical Fiber Communication Conference and Exhibition (OFC). Meanwhile, they used an experimental setup to show how the predictive touch response enhances human-to-machine applications.
Using Predictive Touch Response to Interact With Machines
Human-to-machines applications often require short network response time, depending on how dynamic the interaction is. It could be as quick as one millisecond.
“These response times impose a limit on how far apart humans and machines can be placed,” said Wong. “Hence, solutions to decouple this distance from the network response time is critical to realizing the Tactile Internet.”
To address this issue, the researchers trained a reinforcement learning algorithm to predictive the appropriate feedback.
Before the real feedback is known, the module, called the Event-based HAptic SAmple Forecast (EHASAF), uses a neural network to forecast the material touched. After identifying the material, the unit adapts and updates its probability distribution to select the appropriate feedback.
Expectedly, this can speed up the haptic feedback from human-to-machine Interaction, potentially paving the way for Tactile Internet.
For their test, the team combined the EHASAF module with a pair of virtual reality gloves. The gloves contain sensors on the wrists and fingers to detect a virtual ball. Upon touching the ball, the module went through feedback options to generate until it resolves the actual material of the chosen ball.
Currently, the feedback options include four materials, and the EHASAF module’s prediction accuracy is 97 percent. However, the researchers expressed plans to improve prediction accuracy across a broader range of materials.
