ERICA: Enabling Real-time Mistake Detection & Corrective Feedback on Free-Weights Exercise




We present ERICA, a digital personal trainer for users performing free weights exercises, with two key differentiators: (a) First, unlike prior approaches that either require multiple on-body wearables or specialized infrastructural sensing, ERICA uses a single in-ear “earable" device (piggybacking on a form factor routinely used by millions of gym-goers) and a simple inertial sensor mounted on each weight equipment; (b) Second, unlike prior work that focuses primarily on quantifying a workout, ERICA additionally identifies a variety of fine-grained exercising mistakes and delivers real-time, in-situ corrective instructions.


Our technology supports fine-grained tracking of an individual’s exercise via a combination of (a) sensors on personal wireless ear-worn devices (‘earables’) and (b) inexpensive inertial sensors attached to non-personal exercise equipment. The choice of the earable is key to the practical appeal of our technology: (i) it is the most popular wearable device and widely used during workouts, thereby significantly lowering the adoption barrier, and (ii) it can directly support audio feedback to individuals. Our technology features the key properties below: Dynamic multi-user, multi-equipment support Unobtrusive, low-cost instrumentation Fine-grained mistake detection Timely and moderate feedback Real-world studies show that ERICA can: Robustly handle scenarios with multiple (even 15) concurrently exercising users Identify the exercise performed with 96% accuracy Identify 94.4% of mistakes during the first 5 repetitions The resulting feedback is favored by 78% of users and results in 10% less repetition-level mistakes


This technology can be used in a “Digital Personal Trainer” application that performs advanced analytics to provide personalized real-time feedback. Such an application can reveal workout issues such as Is the user making mistakes while performing exercises? For example, is the user performing full range of motion of lateral raises exercise?, Is the user swinging his body while doing bicep curls? Is the user training the same set of muscles repeatedly across different sessions? The application can then suggest more variety of exercises and also specific areas that a user need to improve. The proposed technology can be employed in the Wellness & Healthcare industry.


Using such simple, cost-effective and unobtrusive approach that integrates quantified fitness monitoring capabilities: Gym-goers will be able to track their exercise routine (exercises, number of sets/repetitions, quality of execution), progress and performance. They can obtain personalized coaching from a “virtual trainer”, which identifies mistake in exercise execution and provides corrective feedback. Gyms and activity centres can offer their users valuable personalized feedback at low cost, thereby increasing customer engagement and increasing retention. Mobile and wearable App developers can integrate gym activity into upcoming wellness and healthy living Apps.