Modalities
Input modalities
Because our product’s main purpose is to give feedback to the user during exercise, input modalities are not as important as output modalities. However, there must be some way for user to control the device, so input modalities need to be shortly discussed too. Since the training clothes themselves will be accessed through the companion app installed on a user’s smartphone, the modalities are restricted to those that the phone offers.
The aim is that there is no need for the user to manipulate the device during the exercise. Before the exercise it should be enough to choose what sport one is going to do and, possibly, what are the goals for this particular exercise. For these kinds of things regular touch input is quite sufficient. If there was a need to communicate with the device during the exercise, then some kind of voice control should be considered. If, for example, the app gave some suggestions that needed a response from the user, a short answer (yes/no) should be enough, in order not to divert the user’s attention from the task at hand.
Output modalities
There are very few sports in which the athlete can look at the screen of their phone while exercising. Therefore, the most important output modality cannot be visual. Our product must be able to give feedback to the user without using the screen of the phone. In the future, some kind of smart glasses might be popular enough to be feasible as the channel of visual output. However, unless visual channel can bring significant advantage over other modalities, our main modality will be auditory.
Apart from being easier to use in sports situations, auditory feedback has some other benefits compared to visual feedback. For example, it is more salient, which makes it a good choice when transferred information has high priority. In addition to that, auditory modality, unlike visual, is not dependent on the direction of attention. However, its disadvantage is that it may steal attention from more important things, which may be risky in some contexts. (Cao et al. 2010, 65–66)
Tactile feedback would also be an interesting possibility for this kind of product. However, since it can be intrusive, and since we didn’t have time to conduct user research on the acceptability of this kind of stimuli in the context of sports equipment, it is presented here only as an interesting prospect for the future development of this product.
Technology
What is the technology behind the product?
The product works by using surface electromyography (EMG) technology. EMG technology records the movement of the muscles by tracking the electrical bursts the muscle generates when contracting, which propagates trough adjacent tissue and bone and can then be recorded from neighboring skin areas. Muscle contraction process starts in the brains motor cortex from where neural activity signals to the spinal cord information about the movement which is sent via motor neurons to the relevant muscle. The motor neurons innervate the muscle at the neuromuscular junction which inturn releases Calcium ions in the muscle. This creates a mechanical charge in the tension of the muscle and depolarization, the difference in current can be detected with EMG. (Farnsworth 2018, Johns Hopkins Medicine)
EMG activity is linearly related to the amount of muscle contraction and number of contracted muscles. In essence, the stronger the contraction and the higher the number of muscles activated, the higher the measured voltage amplitude. (Farnsworth, 2018)
Traditionally EMG technology is used in clinical tests with EMG electrodes which are small needles that go through the skin into the muscle (URMC). These tests are usually performed within hospitals or training facilities with large machines and wires going from the electrodes to the machine. It is usually used to study deep muscles or muscles that have small cross-sectional area. The main advantages the intramuscular EMC has is that it is able to selectively detect EMC signals during static and dynamic conditions and simultaneously minimize crosstalk (Péter et al. 2019).
It is also possible to acquire EMG signals non-invasively by placing the electrodes on the skin surface with surface EMC (sEMC) electrodes. Invasive electrodes (or electrodes with needles) are not possible in our use case, so this is why we’ll be using surface EMG technology or sEMC. This has been proven to be a valid method to track muscle fatigue as its results are reasonable and easily recognizable and the method to analyze the results requires very small computational weight which enables it to be used in low-cost systems (del Toro et al. 2019, Petér et al 2019, Perry 1981). This is one of the major reasons why we believe this technology fits perfectly in our use case.
The benefits of the sEMC technology is that it is quite easy to start using and it still collects accurate enough data to make meaningful conclusions and recommendations on. It is simple to analyse as the output is trivial and it doesn’t require huge amounts of computer power to analyse. (Garcia & Viera 2011) It is possible to get the data in realtime so that user sees the results simultaneously as working out. In addition, the electrodes are light and durable so they will endure wear and tear of users.
How is the technology used?
We will use sEMC in our product by imbedding the electrodes into a piece of clothing. The placing of the sensors is not extremely precise because they are able to collect signals from a large area, but they need to be small enough that it doesn’t affect the usability of the cloth. We would suggest that the sensors are in a mesh within the fabric in the critical areas and connect the mesh to a removable device that collects and sends the data to the companion application where it can analyse the data and show the results to the user.
We will use sEMC in our product by imbedding the electrodes into a piece of clothing. It is not clear according to the research what is the best place for the electrodes to collect most accurate and clean data. It is suggested to place the electrodes on top off bone to not record adjacent muscles signals (Blanc & Dimanico 2010) or in between the motor unit and the tendon insertion of the muscle (Jamal 2012). Moreover the need for a reference electrode is unclear (Blanc & Dimanico 2010, Jamal 2012). The electrodes are able to collect signals from a large area, but they need to be small enough that it doesn’t affect the usability of the cloth. We would suggest that the sensors are in a mesh within the fabric in the critical areas so that they are durable enough. The electrode mesh connects to a removable device that collects and sends the data to the companion application where it can analyse the data and show the results to the user.
The electrodes collect the EMC activity data in a time-series. From the data, we can see how long the exercise lasted, what muscles were activated and how much. This will then be shown to the user in an understandable way. The artificial intelligence will suggest different exercises to activate wanted muscles in a more in-depth level, say what to focus on during the sets to get more muscle activation and better trajectories, and how long should the user rest and what the next workout should be. In addition the AI will clean the data from adjacent or inactive muscle signals.
Future consideration and competition
It is clear that sEMC is the technology of the future when it comes to tracking muscle activation. It has been proven that the results are reliable enough for meaningful conclusions (Péter, 2019). Biggest question at the moment are:
- How affordable the technology can become,
- How can the resulting data be used in a truly efficient and useful way,
- Do the results give a big enough competitive edge.
As in almost every technology, it will inevitably become more efficient to produce and the parts will become cheaper. This will be a big factor on when the technology will become mainstream and breakthrough to the normal customer segment. Another big question is what kind of insights can be made from the data. First obvious use is during the workouts but in the future it might be possible to use the data in everyday life, such as posture checking/lower back pain prevention for office workers (Neblett 2016) or for rehabilitation in injured people. Third major question in the future of the technology is that this technology is already available and in use in some major professional teams, why hasn’t it diffused more and spread into other teams? Major reason is that the there are so few real companies using this technology but does the product give good enough competitive advantage compared to not using it at all.
There have been some startups that have tried to use the technology but they haven’t understood the technology’s limitations enough to make sellable products. The biggest companies at the moment using this technology are Athos from USA founded in 2012 (Athos) and a Finnish company Myontec (Myontec). Athos has been more active during the time of founding the company and has some major partners and customers, such as U.S. Air Force and NFL Team Portland Timbers. Their products include a shirt and leggings for men and women, paired with an app that gives information about training and personalized training programs. (Athos) Myontecs product line offers shorts that collect the data and a MCell that sends the data from the shorts to be analyzed (Myontec). Myontec is quite small company and is searching for funding. This could be a real possibility for Polar to acquire this company if the technology seems potential enough.
An Example User Task
Donald has just started to do weight training. He wants to get better at it, but cannot afford a personal trainer, so he reads blogs and discussion forums as well as watches videos from the internet to learn correct ways of doing things. One day he reads someone recommending a product that includes training clothes that track muscle activity and a phone application that acts as a “personal trainer” based on the information gathered using the clothes. Because the product’s basic functionalities don’t require him to have a monthly subscription (like a personal trainer would do), he decides to give it a shot, after reading a couple of other positive reviews on it, of course.
When the product arrives, Donald can’t wait to try it out. The next day when he comes from school, he dons the training pants and shirt, connects them to the companion app and puts on his Bluetooth earbuds, so he can listen to his training playlist while at the gym. He also tells the application what kind of sport he is going to do.
When he arrives at the gym, he starts by doing bicep curls. Then he moves on to squats. During the first set the application informs him through the earbuds that his right calf is doing more work than the left and asks if he wants to see any suggestions related to that. He has learned from the internet and from his experience that he is prone to a certain wrong posture, which he knows how to quickly fix, so he replies “no”. During the third set, when he is already a bit weary, the application this time tells that his right glute is doing more work than the left and again asks if Donald wants to see some suggestion. This time he has no idea what is causing the problem so he replies “Yes”.
Because the application knows that Donald cannot use his phone during the exercise, it doesn’t offer its suggestions right away, but waits until he has finished his set. During the break between the third and the fourth set, Donald takes his phone from the armband and watches a one minute video describing his error and suggesting a correction. Donald knows he cannot get to the bottom of the problem in this short time, so he saves for later a more in-depth video suggested by the app as well and continues his exercise making his best effort to avoid the error based on the instructions given in the short video.
After workout Donald goes back home, takes a shower and eats a healthy meal. He starts to read a book when suddenly his phone vibrates and the application notifies him about the more in-depth video he wanted to watch later. Donald pauses his reading, takes his phone and changes the application’s settings so that it doesn’t remind of the videos after the training. He is motivated enough to remember to watch them himself when he feels like so, so in this regard he doesn’t need the inciting of his “personal trainer”. He reads the book for an hour or so and goes to sleep.
Team work evaluation
Most of us were nearly always present in the workshops, so the workload for tasks performed there is somewhat equal. Here are our points and work item highlights:
Erkka Juhaninmäki 5p
- Wrote most of the blog posts
- Participated most actively in the exercise workshops
- Made the prototype/product brochure
Esa Niemi 5p
- Conducted most of the interviews
- Made the video presentation and voice over
- Improved the blog posts based on course staff feedback
Anton Pikkupeura 5p
- Edited the video (good job Anton 👍)
- Formed personas
- Researched the product technologies
Sources
Athos, https://shop.liveathos.com
Blanc, Y. & Dimanico, U. (2010). Electrode Placement in Surface Electromyography (sEMG) “Minimal Crosstalk Area” (MCA), The Open Rehabilitation Journal, https://benthamopen.com/contents/pdf/TOREHJ/TOREHJ-3-110.pdf
Cao, Y., Theune, M. & Nijholt, A. (2010). Cognitive-Aware Modality Allocation in Intelligent Multimodal Information Presentation. In L. Shao, C. Shan, J. Luo & M. Etoh (Eds.), Multimedia Interaction and Intelligent User Interfaces: Principles, Methods and Applications. London: Springer, 61–83.
Cavalcanti Garcia, M. A. & Vieira, M. (2011) Surface electromyography: Why, when and how to use it, Revista Andaluza de Medicina del Deporte, https://www.elsevier.es/es-revista-revista-andaluza-medicina-del-deporte-284-articulo-surface-electromyography-why-when-how-X1888754611201253
Fuentes del Toro, S., Santos-Cuadros, S., Olmeda, E., Álvarez-Caldas, C., Diaz, V., San Román, J. L., (2009). Is the Use of a Low-Cost sEMG Sensor Valid to Measure Muscle Fatigue?, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6679263/
Farnsworth, B. (2018) What Is EMG (Electromyography) and How Does It Work?, https://imotions.com/blog/electromyography-101/
Jamal, M. Z., (2012) Signal Acquisition Using Surface EMG and Circuit Design Considerations for Robotic Prosthesis, Computational Intelligence in Electromyography Analysis – A Perspective on Current Applications and Future Challenges
Johns Hopkins Medicine, Electromyography (EMG), https://www.hopkinsmedicine.org/health/treatment-tests-and-therapies/electromyography-emg
Myontec, https://www.myontec.com
Neblett, R. (2016). Surface Electromyographic (SEMG) Biofeedback for Chronic Low Back Pain, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4934580/
Perry, J., Schmidt Easterday, C., Antonelli, D. (1981). Surface Versus Intramuscular Electrodes for Electromyography of Superficial and Deep Muscles, https://academic.oup.com/ptj/article-abstract/61/1/7/2727211
Péter, A., Andersson, E., Hegyi, A., Finni, T., Tarassova, O., Cronin, N., Grundström, H., Arndt, A., (2019). Comparing Surface and Fine-Wire Electromyography Activity of Lower Leg Muscles at Different Walking Speeds, https://www.frontiersin.org/articles/10.3389/fphys.2019.01283/full
Polar Bears 