[Legacy Report] IEEE UP Section Seminar on Applications of Machine Learning in Haptic Perception
#IEEE
#UP
#Section
#Seminar
Haptic data transmission being bidirectional in nature, can tolerate a communication delay of only upto 20ms. Hence, only a small number of samples can be incorporated in a packet, which leads to a high packet rate, causing network congestion. Thus, it is required to reduce the packet rate while maintaining the quality of the perception. To achieve this, a kinesthetic force signal may be sampled at points where the relative difference exceeds the JND (just noticeable difference). It is required to study and identify different adaptive sampling strategies for haptic signals. We first seek to identify good adaptive sampling strategies for haptic signals. Our approach relies on experiments wherein we record the response of several users to haptic stimuli. We then learn different classifiers to predict the user response based on a variety of causal signal features. The classifiers that have good prediction accuracy serve as possible candidates to be used in adaptive sampling. We compare the resultant adaptive samplers based on their rate-distortion tradeoff using synthetic as well as natural data. We also provide data driven approaches to quantify the perceptual limitations of users.
Add Event to Calendar
