or Connect
EpicSki › The Barking Bear Forums › Ski Training and Pro Forums › Ski Instruction & Coaching › Technology for Skiing Technique and Instruction
New Posts  All Forums:Forum Nav:

Technology for Skiing Technique and Instruction

post #1 of 46
Thread Starter 

Hi people!


I'm looking for some feedback and/or signs of interest for some ski technology a couple of guys and I have developed. We've built a sensor unit that retrofits into ski boots and measures the pressure distribution and motion of the skis during a run. The sensor unit connects to a boot-mounted unit which wirelessly communicates with the skier's smartphone. Here's a picture of our current prototype:


Here's a short clip showing some of the raw data we collect:


This technology is designed to help skiers learn about their technique, and to give instructors and coaches precise information which can be used in their lessons. We've tested our prototypes with a few highly competitive skiers and instructors in the Alps, and have received a positive response so far!


We're now wondering if this technology is of any interest to the skiing community at large, so please do leave us some feedback, and feel free to ask any questions! Thank you!


More information about the technology:
The sensor unit has numerous pressure sensors, allowing us to determine the pressure distribution between the left and right feet, distinguish between heel and toe pressure etc (it's pretty awesome). Here's a quick demonstration using a heat map:


We're also developing a mobile application that analyses the measured data and provides feedback about the skier's technique. For instance, we can prompt the skier (in real-time!) to put their weight forward when we detect them leaning too far backwards. Over time, the app can generate a profile for the skier, which can be used to suggest training goals.


We are also looking into gamifying ski drills to make them both more quantifiable and enjoyable. We could provide simple goals such as "make 10 turns" and track their progress using through the sensors. The app could then score the drill based on a suitable metric, have leaderboards and friendly competitions to encourage doing drills!


We are very excited about this technology and its potential, and we hope you are too! Please let us know what you think, thank you!

post #2 of 46

The golf industry is far ahead of the ski industry in use of sensor technology to aid movement analysis and instruction. You might look at that as a model for determining the consumer market potential of this technology and your plans for the future. There's a small market for this. It starts with high end instructors and then works it's way down to the consumer "geek" level market as the price drops. Looking at the first demo clip, the data generated immediately prompted me to go "so what?". Take a close look at the Sky Pro. That product does a great job of turning the raw data into conclusions about swing mechanics and recommending technique changes. But it is limited because it can not analyze results (i.e. ball flight). In golf, sometimes inefficient movements can have acceptable outcomes (e.g. Jim Furyk's swing). With skiing, it is hard to even define what "ideal results" should be (and hard for these kinds of sensors to capture the path of the skis in the snow). because intent and conditions are so variable. But I would bet that if you captured data from high end skiers and compared it with data from recreational skiers with typical technique issues that it would be fairly easy to create a Sky Pro kind of product for skiers. I suspect speed and acceleration sensors would be critical additions. A measurement of slope pitch could be useful if you could figure out how to accurately get it. One definite for downhill skiing is to have lateral/horizontal pressure sensors to pick up foot twisting. The second demo does a better job of translating raw data into a presentation of movement. Look at gears golf. That's a technology that gets much closer to analysis that a recreational skier can understand and use.


As an instructor, the bottom line for me is that the technology could improve my teaching. My experience as a student of golf has been that at least 90% of the time, a coach can figure out issues far faster without using the technology. We don't need fancy sensors in someone's boots to fix a "Z" turner. I'm not worried that a fancy app is going to be very successful at getting many "Z" turners to fix themselves and put me out of business (although personally that would be great). But I could use such a tool to test, train and validate my own "eyeball" observations. I could also use it to enhance my own technical knowledge of the sport.


The bottom line for market potential is that with the prices of a private lesson at a major resort exceeding several hundred dollars, there will definitely be enough people who will want to try a product like this to make getting into this market a good idea.

post #3 of 46


... For instance, we can prompt the skier (in real-time!) to put their weight forward when we detect them leaning too far backwards...!

Exactly, in real time, must be organized feedback with the athlete. What's the use of looking at the pictures on the smartphone, when it was all over. Since a skier has a bit of free senses while driving, it is desirable to put the information into his ears. Examples of such implementation are: http://www.advancedracingcomputers.com/vlink. html www.forwardski.com. I wish you every success!

post #4 of 46
Thread Starter 

Thank you for your response and feedback!


I must admit that the first video does very little to demonstrate the capabilities of our technology. We actually do have software that performs technique analysis! Our algorithms are already capable of detecting simple mistakes such as wrong weight distribution and leaning too far back and make real-time suggestions to correct them.


We think the raw data synchronised with a video will be useful for in-depth technique analysis, probably for racers and other highly competitive skiers, but for most other people, we've developed some simpler visualisations such as turn pressure analysis (see image). This version shows the average pressure applied by each foot relative to the estimated path of the skier, but we can also display the pressure distribution for each foot at any point in time.


We have also developed other algorithms to detect possible problems with a skier's technique but as you mentioned, it is very difficult to find the 'ideal' technique. We've experimented with machine learning techniques as a solution to this, but it definitely requires more work.


Thank you for your feedback! Please let us know if you have any more feedback!

post #5 of 46
Thread Starter 

@igoriz Thank you for the feedback!


We agree that real-time technique suggestions will indeed be very useful, and we're certainly working on getting as much as possible real-time. However, we also believe some racers may be looking for very small or very specific features in their technique (e.g. perfect weight distribution on jumps) which may be better done after the run with a coach when there's time for visualisation, thought and reflection!

post #6 of 46

Isn't the human body already wired to sense pressure changes under the foot in real time? Being taught by an instructor or coach to sense what is happening in real time seems to me to be hard to beat.


That said, I'm sure there will be plenty of tech junkie ski nerds with money to throw at it. There are a lot of companion apps for sensor and information technologies with a market life cycle that will eventually reach a negative tipping point. In other words, if you can make money with this, make all that you can while you can before the niche novelty that we see with similar technologies wears off.


One issue I would bring up regarding coach and instructor use is the lack of transfer-ability from one skier to another regarding its placement inside the boot.


Rusty makes a helpful point in identifying potential from collateral uses already in mid-market. It may also help to research the success of all the other ski apps that sense and display max, min and average speed, direction, elevation changes, g-forces, etc. that also require no further equipment acquisition than the app itself. The novelty factor with these apps may have proven to be temporary.


It would be interesting to see the product and business model screened on the Shark Tank. Even a failure there will provide an initial boost. While its invest-ability may be a positive indicator, non-invest-ability is not necessarily a precursor to failure.


There are a lot of hidden factors that determine whether a product takes off or not, such as word of mouth support so I wouldn't be surprised to see your product become successful.  

post #7 of 46
Thread Starter 

@Rich666, thank you for your feedback and optimism! Our prototypes work together with the smartphone, so it can most definitely do everything ski apps can and more.


I agree that the skier can certainly feel the pressure he/she is applying during skiing, but the instructor watching the skier usually has to employ some form of informed estimation. We hope our technology can reduce the amount of guesswork and complement the skills of an instructor!


Another potential advantage is that the technology will be able to stay with the skier for the entire duration of skiing, monitoring every move and every turn. I imagine over a long day of skiing, a skier's technique will continuously change to avoid straining the same muscles, something we might be able to observe with the technology. I'm personally very interested to see this kind of data, especially for endurance events like cross-country, and even in other sports such as marathon running, distance cycling etc.


Our current sensor prototypes are designed as thin footbeds/inserts which slip under the boots' liners. They are fairly easy to install and remove with most boots, and we haven't had any problems during our testing trip in the Alps!

post #8 of 46

Could gps measurements in g-force and turn radius be compared with pressure changes under the foot to differentiate down pressure from the skier with up pressure from the centrifugal force of the turn?

This is something I would find even more interesting but perhaps gps would not be accurate enough for the outcome sought.


Also, you mentioned that the pressure sensor data is being displayed on a heat map. Are you measuring pressure or the heat from increased pressure? I know pressure can change at a much quicker rate than temperature so I feel there is something I am not understanding completely. I think it could raise similar inquiries with others who, like me, do not have a science background.


I am sure it has run across your mind's long ago, but have you yet attempted to gain organizational backing such as an adopted program integration with one of the known teaching or coaching organizations? If so, I would be curious as to the results.  I see a quicker potential for application with some of the online remote coaching services. As the cost for privates continue to sky rocket, I think we will see more people considering these sites as a more cost effective means of improving their skiig. Emailing video for movement analysis is a key factor to these services and including the data that your product measures may be considered a very relevant value add. Selling the product directly through these sites may be an avenue to consider as much as I am sure you must have considered already as well. As a matter of fact, it may give these remote coaching services more edge for their own marketing.

post #9 of 46
Thread Starter 

@Rich666, that is an interesting question! We can use the position and velocity outputs from the GPS receiver and combine them with the angular velocity from our gyroscopes to estimate the centripetal force the skier has to apply in order to maintain the turn/circular motion. Our accelerometers also provide related information. Furthermore, we could also estimate the weight of the skier when he/she is stationary. There will undoubtedly be some complications with noise, the actual acceleration of the skier down the slope etc., but it seems quite possible to maintain estimates of the components of the force applied by the skier. I'm interested to try this!


Sorry about the confusion regarding the heat map, we're measuring the pressure applied by skier. The amount of pressure is represented in the as colours. In this case, green is low pressure, red is high pressure. Now that I think about it, the name "heat map" isn't exactly intuitive.


We have thought about working with many people to get this technology started! Collaboration with instructors is definitely of great interest to us, since this is where we think our technology will flourish. We've also thought about many potential features like letting users compare their technique/stats/data against a professional skier's and remote coaching. We believe this technology could make a difference in how skiers learn and how instructors teach in the future. But first, we'll have to find out if the skiing community is receptive of such technology!

post #10 of 46
I hereby volunteer to field test this free of charge! :-) I think the "heat map" could prove to be interesting from a bio-mech pov...

post #11 of 46
Originally Posted by zentune View Post

I hereby volunteer to field test this free of charge! :-) I think the "heat map" could prove to be interesting from a bio-mech pov...



Well Zenny, giving some of these out to the right people such as ski instructors who often post on ski forums, very well could be a small piece of a successful and active marketing plan. So, I second Zenny in this regard! Using some of those field tests to acquire footage for a youtube demo videos and even perhaps animated graphic demos both to put on a product website may help. The product usage is process oriented enough to warrant some type of visual demo. Getting booths at any upcoming ski shows, have it be one of those free prizes they throw out at the beginning of every annual Warren Miller showing, having presence at a demo day or two and doing what it would take to get a few ski academy coaches using it on their athletes. I suppose, like anything similar, you are going to have to gain a certain amount of buy-in before you start production which can be expensive and starts the launch sequence in stages, each of which, you need to be ready for. New product lines can be a tough business.

post #12 of 46
Thread Starter 

@zentune, thank you very much for volunteering! It's always great to have people enthusiastic about our technology, it means a lot to us. We've only just completed a field test recently in the Alps, and we're looking at developing the next revision if we receive enough interest and support. I'll definitely remember your offer if/when we are ready for beta testers!


@Rich666, thank you for the suggestions, we'll definitely look into them! We're trying hard to gauge the level of interest and possibly gather the support from the skiing community so we can continue developing the technology. We've actually had the opportunity to meet and test with some highly competitive skiers when we were in the Alps, and have received a very positive response. We now hope to reach out to the wider skiing community, and we'll be immensely appreciative if anyone were to help spread the word to colleagues, friends and family!

post #13 of 46

This is interesting. Certainly you could profile a few drills executed perfectly and compare people's performance on any given drill run and provide feedback.


I developed something like this, with some guys, long time ago, for testing suspension elements of a car. there were 4 plates for the wheels which were shaking the car and some sensors measured the response harmonics and stuff and via some very complicated wizard-level Fourier math figured out any problems with the suspension elements.


You could use something like that to compare to a benchmark for drills and tell them feedback at the bottom of each run and THAT instant feedback is priceless, in todays' partnerless society.


for an effective system though, you must couple that with a simple, effective and measurable instruction system and a website to manage it all, track progress and mediate dialog with remote instructors and the results could be awesome. I have some ideas there as well...



P.s. couple it all with headphones and instant feedback... yummy. How do we mix this over Metallica? That will be the hard part. Or better yet, how do we get Jeremy's voice: "Lean forward, you blathering idiot"... when they crank the feedback level up to "spank me"!

Edited by razie - 11/2/15 at 7:31pm
post #14 of 46
Thread Starter 

@razie, thank you for the feedback! We have similar thoughts, and we're working hard to design an easy-to-use system that ties everything together. We are keen on working with instructors to develop a rigourous lesson plan, supplemented by a comprehensive set of drills, all monitored using our sensors to give feedback and encouragement. If you have any ideas or suggestions, please do let us know!


For technique analysis, we're looking into a whole range of analysis methods, from the more traditional like Fourier (we might be able to distinguish whether you're riding on powder or ice) to modern machine learning and even bleeding-edge deep learning.


Selectable feedback voices sounds like a great way to make everything more fun! I'm definitely keen on giving that a go. Thanks for the suggestion!

post #15 of 46

Sounds very interesting!! You could record the results from skiers like Jf and Rogan doing various speeds, turn shapes, pitch, terrain and analyze the differences. You could campare your own to theirs. You could record your results on different skis: models, lengths. Wow, so much fun!


Count me in for a beta program.

post #16 of 46
Thread Starter 

@Blizzard Thank you for your interest, we really appreciate it! Being able to compare our technique against a professional's is definitely in our vision! We imagine it'll be super helpful to learn exactly when, where and how to apply the pressure on a turn to achieve a desired result, and to learn by emulation in general.

post #17 of 46

This is very interesting. I'll admit that the first few posts confused the hell out of me, but seeing the pressure analysis was really cool. Count me in as well for any beta testing. Whats more, when you're closer to production, I'd like to mention this to my supervisors at Whistler. I'm sure they'd be very interested, even if they used it just for instructor training.

post #18 of 46
Thread Starter 

@FlyingFish Thank you very much for your interest and support! I'll have a chat with the team, and we'll definitely be in contact once we've figured out how we can offer a beta programme. Out of curiosity, which parts of the previous posts did you find confusing?

post #19 of 46

We're also developing a mobile application that analyses the measured data and provides feedback about the skier's technique. For instance, we can prompt the skier (in real-time!) to put their weight forward when we detect them leaning too far backwards. Over time, the app can generate a profile for the skier, which can be used to suggest training goals.


How are you planning to prompt the skier real time?


I can't look at my smart phone and ski along them at the same time.  If I'm watching them and not the phone/app, then I don't need the app.  Are you planning some sort of communication device, like something as simple as having the student wear a bluetooth device and then call them so you can give them feed back?


I find this interesting and compelling, but my concern is it would be for a narrow market.  Ease of use will be one of the biggest selling features.



post #20 of 46
Thread Starter 

Thank you for your interest! We've currently tested real-time feedback through earphones with voice prompts. We've also received suggestions to use haptic feedback through smart watches, and visible prompts on head mounted displays, so we'll definitely be trying those out soon!


We're designing our hardware and smartphone app to be extremely easy to use. All the raw data will be analysed by our algorithms and summarised into both actionable feedback and technique metrics. Besides the real-time feedback, the user will also be able to view the full analysis on the app after the run, include the pressure analysis and other factoids like jumps, spins etc. More advanced users, instructors and coaches can get access to the raw data that can be automatically synchronised with a video taken using an iPad for in-depth technique analysis.

post #21 of 46

For those who wish to develop their technique as fast as possible, I recommend electrical shock feedback. Speaking on selling features, your'e killing two birds with one stone if you have schizophrenia.

post #22 of 46
Thread Starter 

On a somewhat similar note, some people suggested we use electrodes to "remote control" the skier's body to teach them how to ski, a little bit like what's shown in this talk:



That technology is just a little less mature than we'd like though.

post #23 of 46
Well, it sounds more mature than some skiers I know....
post #24 of 46
Thread Starter 

Thank you guys for all the feedback so far!


Here's as short clip of our test with ski cross world champion Filip Flisar:


We can very clearly see the timing and force applied for his jumps, landings and the final turn at the end of the video.


Apologies for using the plots again, we're still working on improving the data visualisation!

post #25 of 46

motionmetrics, I've been waiting for such a platform to come into existence! It looks like a great objective indicator of performance. 


I'm interested in how you address effect vs cause, and detection vs correction. For example, pretend the sensors determine that a skier is inconsistently or aggressively pressuring certain parts of the ski throughout the turn. Some causes could include: 

  • The skier is bent over one hinge point (e.g. waist)
  • The skier is standing rigid, with lots of tension in the legs (and other joints)
  • The skier's core isn't engaged
  • Something visibly imbalanced in the stance
  • Non-sharp skis hitting a patch of ice



Like rusty mentioned, most of these types of root causes are easily detected by an instructor. Do you expect the app you build to be able to detect root causes? 


You may have explored this already, but I'd recommend reaching out to indoor ski centres to sell them on the technology. Since the indoor centre is a closed environment, skiers could see the live data on-screen as they practice. A live feed will help the skiers immediately connect their motor performance to the outcomes on-screen. Immediate feedback is vastly more effective for creating change than delayed feedback. 


Another obvious environment is the race course. Would be great if you could enable coaches to easily create a visual overlay of their data on the race course, and compare across their race team. 


Keep us posted on what happens!

post #26 of 46

As an Instructor using Video a lot for movement analysis, and access to several instructors, I'd be happy to be tester. We could video runs with your units in place and maybe you can use the data and video, to assist with understanding the pressures and movements.


Would you also be interested in using similar technology for Archers? force pressure plates, for weight distribution, shot execution, movement during shot execution, etc? Put me on the list for that as well!



post #27 of 46
Thread Starter 
@Metaphor_, thank you for your interest!
You're right to say that our current system will only be able to detect certain symptoms, like in your example. We're working on algorithms that may be able to eliminate some of the possible causes with additional data we are able to collect, but in the short term, we do not expect to be able to conclusively recognise the cause for all symptoms. If the skier is with an instructor or coach, we will be able to flag up the symptom and display it along with a synchronised video (if available) for analysis. If not, we think a good way will be to prompt the user to consider a list of possible causes and/or to go through a list of tests to further eliminate the causes.
About indoor ski centres, thank you for the suggestion! It might be a great avenue for us to test our technology. Of course, our system actually works with a smartphone, and provides real-time audio feedback through earphones anyway, so the skier can still receive instant feedback up in the mountains!
post #28 of 46
Thread Starter 

@dchan, thank you for your interest too! I'll definitely be in contact with you when we've figured out our beta programme (hopefully in the coming weeks).


We're actually also looking into supporting other sports. Our core technology is very versatile, and we've tested with runners and cyclists too, but in terms of real-time feedback and technique analysis, we're currently committed to focusing on skiing! I can certainly see the applications in archery, shooting (I was a competitive shooter) and other related sports, and it'll be fantastic if we could see which other sports are interested in this technology. We're seriously considering having a developers' programme for people who want to test this technology with other sports!

post #29 of 46

This is very interesting.   I'd love to test it as well :).


As you've mentioned, there are two broad ways in which this can be used, and different approaches will be needed to optimize each.   The first, and most obvious, is post-run data analysis, which you've displayed above.  The second, as you described, is real-time audio feedback to obtain corrections during the run.  The latter is particularly intriguing, but a skiing implementation of this presents some interesting challenges, because the proper endpoints vary continuously through the turn.    Consider, for instance, fore-aft balance.  In elite skiers, the weight distribution is most forward at and and after the apex, and most back when they flex to release at the transition (this has been confirmed by kinematic studies).  So a fore-aft distribution that is appropriate at at the transition is going to be too far back at the apex.  So how will your instruments "know" where the skier is in the turn?  One obvious way would be to sense the ski angulation, and its rate of change.   More broadly, I think if you're going to be producing this for downhill skiing, and want it to be a powerful technique-monitoring and analysis device, ski-angle sensing will be important.


Another technique issue is avoiding an A-frame (i.e., making sure your inside ski is angulated at least as much as your outside ski).   You could build in a variable-volume warning sound to alert the skier when this isn't happening -- as the outside angle increasingly exceeds the inside angle, the beeping gets louder   Also, some skiers work on the timing of the tipping of the inside vs. the outside foot.  You could allow the skier to set the device to monitor and give feedback on this.  Likewise, if the skier is working on keeping most of their weight on the outside leg. you could set a warning to go off if the pressure distribution becomes too symmetric. This would vary with the phase of the turn, so you'd need to use your angulation sensors to adjust accordingly.    Rich666 mentioned we already have pressure sensors in our feet.  That's of course correct, but in the midst of the dynamics of a high-performance ski turn, it's easy to over-pressure the inside ski without realizing it.   So a warning could be very useful.  Yet another technique issue is allowing the inside foot to drift too far forward.  You could perhaps monitor for that by seeing when the inside foot has a an overly aft pressure distribution.


Finally, if you could collect not only fore-aft, but also lateral pressure distribution under the foot, this could also be very useful for technique development.  

post #30 of 46

it's hard to make this a generic feedback mechanism. it's a lot easier if it's coupled with an instruction system and an evaluation. I.e. do this drill and it will tell you if you got it or how to improve. An online teaching progression, with clear evaluation and exit criteria would be ideal like... I don't know... mine? :rolleyes

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Ski Instruction & Coaching
EpicSki › The Barking Bear Forums › Ski Training and Pro Forums › Ski Instruction & Coaching › Technology for Skiing Technique and Instruction