Abstract:
"Many began exercising at home after the COVID-19 outbreak due to gym closures and health
concerns. This shift in exercise habits, which led to a rise in interest in at-home exercises, was
caused by a number of factors. When working out at home, many people have a tendency to
assume incorrect posture because they are not receiving expert guidance or assistance.
Numerous health issues as well as decreased workout effectiveness could arise from this.
The bulk of current research on body posture feedback has focused on refining techniques for
accurate body tracking rather than the feedback that comes after body tracking. The feedback
mechanisms that are incorporated into these systems are not given enough consideration. The
majority of previous research has been on the technical aspects of sensor fusion methods and
tracking algorithms in order to deliver dependable, real-time body movement tracking.
However, there are fewer studies that concentrate on employing machine learning (ML) models
to predict joint placements in real-time and offer consumers perceptive feedback based on this
data.
In order to address the problem of poor posture during particular exercises at home, this
research uses augmented reality (AR) and machine learning (ML) to provide real-time
feedback to the reader. This document also includes a brief description of the issue, an analysis
of research gaps, research challenges, and the purpose of the study. The essential proof of the
issue, the research that has already been done and its drawbacks in comparison to the system
that was established, and a synopsis of the author's strategy for enhancing such systems in order
to boost accuracy, efficiency, and user-friendliness are also included."
