Implementation of smartwatch user interface using machine learning based motion recognition

Self Cite

Current consumer wearable devices such as smartwatches mostly rely on touchscreen-based user interfaces. Even though touch-based user interfaces help smartphone users quickly adapt to wearable devices with touchscreens, there exist several limitations. In this paper, we propose a non-touchscreen tactile wearable interface as an alternative to touchscreens on wearable devices. We designed and implemented a joystick-integrated smartwatch prototype to demonstrate our non-touchscreen tactile wearable interface. We iteratively improved and updated our prototype to improve and polish interaction ideas and prototype integration. To show feasibility of our approach, we compared and contrasted form factors of our prototype against the latest nine commercial smartwatches in terms of their dimensions. We also show response time and accuracy of our wearable interface to discuss our rationale for an alternative and usable wearable UI. With the proposed tactile wearable user interface, we believe our approach may serve as a cohesive single interaction device to enable various cross-device interaction scenarios and applications.

Section: Tactile Interactionmentioning

confidence: 99%

Section: Interface Response Speed and Interface Control Accuracymentioning

confidence: 99%

A Non-Touchscreen Tactile Wearable Interface as an Alternative to Touchscreen-Based Wearable Devices

Yoon

Park

2020

Self Cite

“…Feature Extraction. As described in the work by [ 13 ] and [ 14 ], we used similar features to form two different feature sets for our experiment. First feature set includes a total of 84 features; 7 statistical features (mean, standard deviation, max, min, 3 quantiles) for 3 sensors’ 3 axis (x, y, z axis) and the magnitude (m) of the combined axes.…”

Section: Experiments and Exemplar Casesmentioning

confidence: 99%

“…To demonstrate feasibility of our approach, we present two applications (i.e., gesture recognition [ 14 ] and smart factory data collection [ 15 ]) on mobile collocated interactions with wearables from our earlier work.…”

Section: Experiments and Exemplar Casesmentioning

confidence: 99%

Cross-Device Computation Coordination for Mobile Collocated Interactions with Wearables

Yoon

Shin

2019

Self Cite

Mobile devices, wearables and Internet-of-Things are crammed into smaller form factors and batteries, yet they encounter demanding applications such as big data analysis, data mining, machine learning, augmented reality and virtual reality. To meet such high demands in the multi-device ecology, multiple devices should communicate collectively to share computation burdens and stay energy-efficient. In this paper, we present a cross-device computation coordination method for scenarios of mobile collocated interactions with wearables. We formally define a cross-device computation coordination problem and propose a method for solving this problem. Lastly, we demonstrate the feasibility of our approach through experiments and exemplar cases using 12 commercial Android devices with varying computation capabilities.

“…Finally, the logistic regression algorithm is a probabilistic model that estimates the probability of the occurrence of an event using the linear connection of independent variables. Using the algorithm, six motions based on the gyroscope, accelerometer, and linear acceleration are used to improve the user interface [ 21 ].…”

Section: Related Workmentioning

confidence: 99%

Smartwatch User Interface Implementation Using CNN-Based Gesture Pattern Recognition

Kwon

Park

Choi

2018

In recent years, with an increase in the use of smartwatches among wearable devices, various applications for the device have been developed. However, the realization of a user interface is limited by the size and volume of the smartwatch. This study aims to propose a method to classify the user’s gestures without the need of an additional input device to improve the user interface. The smartwatch is equipped with an accelerometer, which collects the data and learns and classifies the gesture pattern using a machine learning algorithm. By incorporating the convolution neural network (CNN) model, the proposed pattern recognition system has become more accurate than the existing model. The performance analysis results show that the proposed pattern recognition system can classify 10 gesture patterns at an accuracy rate of 97.3%.