kiwiwearables-blog1

[kiwiwearables-blog1]

Motion as an alphabet

We move quite a bit each day, however how we break down our movement into sources of input is where we are aiming to go with Kiwi. Think of short movements you make as a keystroke, a swipe of the hand can be a trigger for hello or a rotate to be a click motion.

The challenge like any new source of input is creating an understandable sequence for input similar to the roman alphabet and numerics we know, there exists a range of motions that we perform everyday which can be translated into a series of inputs. These inputs can be used as a shortcut to start your more controller use, for make a chop motion to open music control on your watch, then use buttons or touch moments to change songs.

Another example would be if you see a call come in you can swipe to decline that call.

[kiwiwearables-blog1]

$22 standing desk

A few months ago, I made a $22 standing desk, and have to say that standing is not as bad as I thought it would be. Its basically made with IKEA Lack table, two brackets and a shelf.. see this post for more details

One some days, and mostly after lunch I will be found on the couch next to my desk but overall I stand more than I sit which should account for something.

Kiwi's technology can detect sitting v. standing from either a watch or a phone, using the accelerometer and gyroscope together with our motion recognition engine.. Get in touch if you'd like to learn more.

[kiwiwearables-blog1]

Why activity trackers suck

Motion sensors are prevalent in almost every mobile/wearable device in the market, however their use is still limited to counting steps, and slowly expanding now to detecting a few more motions (swimming is the most popular activity in case you were wondering), followed closely by cycling. Here is a visual representation of motions tracked by some common devices on the market, and an indication of the accuracy level.

We all think activity tracking is a fabulous idea, well maybe everyone but our doctors ... but lets face it, the actual products in hand don’t really turn everyone’s crank just yet.

The idea of counting steps is great, but each motion sensor predicts a relatively random number of steps taken so the benefits today are similar to that from a $1 pedometer and tracking sheet that my first employer used to hand out as part of their “wellness program”.

A typical motion sensor sends 50 values per second from the accelerometer, gyroscope and magnetometer. Below is a quick 2 line explanation on how motion sensors work, just because we get asked this question quite a bit.

There is a minuscule pendulum in a little box, which moves around when the sensor is moved, and approximates values that change on each of the x, y and z axes while compensating for gravity.

What happens next is that the device tries to match the incoming data to a set of pre-populated data for the motion, lets say the ‘typical step' to detect if a step has occurred, and since each device runs a different motion sensor, with a different guess on what your step could be like.. Try doing this ’typical motion’ approach for many more motions e.g. rowing, tennis, baseball, fall detection and the error rate starts to multiply pretty quickly, rendering the detected value quite useless. It turns out that my $1 pedometer and today’s accelerometer/gyroscope combos work on the same principle. Albeit some of today’s activity trackers have great design, awesome connected apps and an improved but still ’typical step’ focused algorithm compared to my $1 pedometer. Options to improve the accuracy and range of motions detected: Use a more expensive, and accurate sensor. Would work pretty well, but would you fork out $1,000 for a super accurate activity tracker. Probably not.  Collect enough data to map and improve motions for a large data set this improves step counting accuracy but what about all the other motions that I make every single day like sitting, standing, walking, running, waving, stretching, … you’ll need to collect a lot of data before that starts working well. Pull a 'Nike Fuel Points', they basically said forget steps and calories, lets just count “fuel points”. Needless to say that introducing another new metric is not the way to go. Bravo to Nike for making that leap, I don’t think many other companies would have the fortitude to make that happen. Our hypothesis on the answer is that  Make it Personal. Each person has a different range of motion, and until software can detect and adapt to a person’s specific range of motions off a template, we’re going to have a tough time moving beyond step counting and low motion complexity sports. Yes, software is eating the world.  More on this in upcoming blog posts. Please also check out our showcase app on Android-wear at bit.ly/kiwiandroid

[kiwiwearables-blog1]

Thoughts on how to get a 3D printer in every home

3D printing isn't directly related to wearables, but we use 3D printing to make the majority of our prototypes. 3D printing is very common in the industrial design industry, and has been for many years. Recently though prices have dropped considerably, leading some people to believe that they will see mass market adoption. Here are a few ideas that might accelerate that adoption. 

Food safe

Most consumer level 3D printers create objects in ABS and PLA. There are variants of these plastics which are food safe. However, the majority of filament used in 3D printers is not food safe. This is especially an issue for kids toys, and anything for the kitchen.

Finished parts right out the of the box

This is slightly more ambitious, but if a consumer (not a hobbyist) is to use a 3D printer, the parts need to be ready to use right out of the box. The reality is that on any 3D printer, especially in the sub $10,000 range, parts made by these machines are quite rough. They need to be sanded, trimmed with an Exato, or exposed to acetone fumes to be smoothed out.

In many cases support material also needs to be removed in an acid bath, or broken off by hand (or with power tools).

  Embedded electronics

If consumers are really going to buy a $2000+ machine, it had better be able to print a pair of headphones, or a TV remote. That’s when the dream of localized manufacturing will be realized.

[kiwiwearables-blog1]

Ensuring an intuitive experience for users

One of the most important items when building an application with a new device is providing the flexibility and control of integrating device. We have found allowing developers to control the modes of the device as a number of states and enable specific streaming states enables more reliable and robust applications to be created.

Having sound and purposeful constraints for developing both software and hardware is paramount. 

Following three simple rules has helped our input:

- providing a simple method to get the output you want

e.g. a simple response like an email when an event happens

- enable a single point of control for sensitivity

e.g. set a range that makes sense, 1 is low and less likely to occur, 10 is high and more likely to

- drag and drop experience

e.g. select the important information to you

[kiwiwearables-blog1]

HackerNest Construct --> Next Level KIWI Apps :)

Hello Everyone, Ashley here :)

Just letting you know that over the past couple of weeks some developers have been building some next level stuff on our platform and I'd like to give them some recognition.

Last weekend at the HackerNest Construct Hackathon 4 teams presented hacks using the Kiwi Move... and they built some awesome stuff!

So please allow me to introduce them:

Team AutoFit

Team AutoFit built an amazing workout app which combined the powers of the the Kiwi Move, Apple iBeacons and the Pebble Smartwatch to give a person an automatically classifying workout log. 

The iBeacons, attached to workout equipment, would tell the Kiwi Move which types of motions to detect (ie. a Move attached to the Chin Up machine puts the device into "Chin Up" detection mode), and the results were output to the Pebble Smartwatch and a companion iOS app.  

Their demo and product were so cool that they presented at the Wearable Wednesday event this past Wednesday.

Let's just say that there's an incredible amount of potential in this app and we're going to be working closely with Ben and Marko to help them get this amazing concept to market!

-------------------------------------------------------------------------

Team Headbanger!

These guys from Guelph! (UofG alumni here) built an awesome headbanging application wherein a player faced off in a headbutting game with a virtual opponent in a location determined by google streetview.  

The player would wear a Kiwi Move on their head, then pick where to face off with their opponent.  The Move detected motions enabling the player to dodge left and right and then headbutt their opponent!

It was like "You, Me, At SEARS!" and you'd be whisked away to a SEARS to headbutt battle.  (hilarious and awesome)

---------------------------------------------

Team ISIS!

These innovators did the world of cooking a huge favour this past weekend!  They managed to get the Kiwi Move to detect whisking, beating, kneading, cutting and other motions which people would use in combination to LEARN HOW TO COOK!  Yes, that's right.

Now all you need to have to learn how to cook isn't some recipe book, ingredients and resilient tastebuds, it's literally a Kiwi Move and an APP! NEXT LEVEL!

---------------------------------------------

Team Myxolidian

The next team, Team Myxolydian, built a really cool gesture enabled midi instrument controller.  It was a monumental task and while everything didn't quite come together, all the pieces were there.  It was something to see!  Being a musician myself I look forward to seeing the completed product, it's definitely something that I'd use!

---------------------------------------------

I'm so impressed with everything that I saw at the Hackathon.  It pleases us at Kiwi to no end to see people using the Kiwi Move and the Platform to build wearable solutions to problems which go so far beyond activity tracking.  

I can't wait to see what people build at the next hackathon :)

[kiwiwearables-blog1]

Just before his "Think Thursday" chat Ashley was interviewed by Bitmaker Labs' Tyler Hackbart!  

It was Tyler's first-ever time interviewing anyone -- We think he did great :)

[kiwiwearables-blog1]

The Kiwi Move

The Kiwi Move is a sleek, non invasive wearable, technology product.  It's designed to suit a variety of human needs, where motion can be optimized to one's benefit.

Packed with advanced micro computing and sensor technology, the Move is designed with a personal customizable voice and gesture interface, offering a better way for people to create technology that is integrated and passive to one's lifestyle and needs. 

The Move works with a Kiwi companion app on your mobile device and makes it easy to take advantage of the variety of apps and plug-ins on the platform.

There are a wide range of applications where the Kiwi Move offers valuable insight, ranging from sports training, personal wellness, lifestyle improvement, connected automation and more.

Kiwi is a pioneer open wearable computing platform providing the tools to build  new and innovative explorations of motion and expand the paradigm of wearable tech. We make it easy for developers to build apps, and give people a platform to select the apps and shape the future of wearable technology. 

[kiwiwearables-blog1]

Q&A for the Kiwi Team

Exactly what is a “multi sensor, internet-connected, open wearable technology device and platform”? 

With a microphone and motion technology, the Kiwi Move can respond to voice commands and gestures to control devices or track activities. The kiwi move device has motion, temperature and sound sensors on board (multi-sensor), these sensors can combine to provide you specific information to help improve your life, for example, if you wanted to create an alert system to notify some one you cared, about a fall, our device could capture the motion of the person falling and if they said "help" this would cause an alert to be sent to the internet to notify those who could assist.

The purpose of the open wearable technology statement is to note that we provide our architecture to developers who wish to create their own applications with our device, the platform notes that we also provide the tools to these developers to ease the process of creating applications with our device and sensor inputs.

How does it relate to my or anyone’s daily life? 

The device could be used as an activity tracker to help you know how much you have moved in your day and set a goal to which you want to aim to meet, the large advantage this device provides is in the realm of biofeedback for improvement on repetitive motion. For example in physiotherapy the device can be used as an exercise tracking and improvement tool.

What human need does it address? 

The human need that is addressed is the desire for continuous improvement and competition to make ourselves better at the activities which we participate in.

How is it used, and how are its uses advantageous?

The device is used by being worn in a variety of locations depending on its use, for example if a person were to use it while they were playing soccer they would attach it to their ankle to determine how many kicks, passes, time spent jogging/running during a game. The advantageous portion is that specific single motions and audio inputs can be identified for each persons desired application, the feedback portion is done through a computer such to provide insight into the captured motion and sound events.

John David

[kiwiwearables-blog1]

Ready to Launch

Its past midnight, and we are working away at putting the final touches on our consumer product launch campaign. Most of us have had little or no sleep for the past few days, and that has pretty much been the story of our lives for the past six months. A story by the way, that all of us wouldn’t exchange for anything else.

We started Kiwi Wearable Technologies six months ago; I personally left a comfortable career in consulting (boardrooms, hotels, airplanes, … repeat) to dive full-time into the exciting and attractive world of entrepreneurship. It has lived up to its promise; I’ve had a great time building an awesome product that has the potential to improve the lives of millions of people.

Things are going well, we understand that our consumers will decide the outcome of our pre-order campaign, however we’re excited about the opportunity for a few reasons 

We’ve made a resolute effort to build a product that people will actually want to wear, and use for many things in their daily lives.

We’ve sold our product to many early adopters, and used their feedback to build a product that people will want to wear for a long time.

We manufactured several versions of our product in the past six months, and have a clear understanding of our road to shipping the product in July 2014.

We’ve got a lot of cool videos of our product in action that we will be launching over the course of the coming days.

We look forward to your support on our campaign to deliver an experience that is easy, useful and engaging. Getting all three exactly right is a tough nut to crack, however I’m confident that we've put our best foot forward, and that this campaign will further steer us in the right direction.

The most important note, I’m proud to work with such an awesome team. John, Ashley, Olivier, Zaki and Mike have all been instrumental in bringing us this far. Our families, friends and members of the thriving Kiwi community have been exceptional in their support for our company and we would not be here without them.

We look forward to your feedback and support in the days to come.

Best Regards,

Ali Nawab (@alinawab)

[kiwiwearables-blog1]

Wearable Tech Predictions for 2014

Hey guys,

I’m Ashley, Chief Marketing Officer at Kiwi Wearable Technologies, and I hope you’ve all celebrated a safe, glorious Christmas/Holiday Season and New Year.  Over the course of the next few days I’m going to release my set of predictions about the evolution of the Wearable Technology space for 2014.  It’s going to be an amazing year (Wearables are HOT! )  and it’s going to be interesting to see in December 2014 just how close/far off I am with these.  That said, let’s get started!  

Prediction Axis #1:  Wearable Technology Competitive Landscape

The great modern technology startup challenge is to align your marketing and your technology development in such a way that you are able to rapidly develop and market a product in parallel.  And neither the marketing nor technology sides of the business can afford to get too far ahead of one another.  Underdevelop and rapidly push out a half-baked product?   You’d better have deep pockets and a sweet crisis management structure in-place or you’re headed down “one-product wonder” street.  Overdevelop and push out too slowly?  You’re probably going to be left behind.  You’re definitely not going to get the traction you need to stay relevant and afloat.

Back in May, when we started Kiwi Wearable Technologies (it seems like so long ago now), there was already quite a bit of competition in the space.  This competition intensified considerably this year, as pre-eminent major players closed large funding rounds (Misfit’s $15M Financing Round, Fitbit’s $43M Financing Round), and made major acquisitions.  (BodyMedia acquired by Jawbone)  These funding rounds and acquisitions are helping these players buy traction to establish their market positions, acquire talent, solidify their intellectual property positions, and assure their future product development plans.

The biggies are coming (if they weren’t already in) and every major technology company is now publicly talking about doing something with wearables.  Apple’s lurking in the bushes, Samsung’s crazy Galaxy Gear watch is out, Google’s probably got an entire warehouse of cool wearable stuff in Project X and even Nissan has a smart watch for god sakes!

The huge challenge for companies in wearable tech in 2014 will be to fund and deliver products which are truly a year ahead of their time… in 6 months.  It’s growing and changing that fast.  Luckily technologists have access to cheap sensors and parts and rapid prototyping in the form of 3D printing and open-source software to make this a bit easier.  Anticipating the market can be akin to looking into a crystal ball sometimes, but at least we have some development models to use as reference (more on this later).

My 2014 Predictions:

We’re going to see some amazing devices this year which will truly blow our minds and captivate us by doing things which were previously considered impossible.  Even more exciting:  what’s gonna be hot in wearables in June 2014 is likely being built in someone’s basement this very moment.

In addition:

-          Wearables will increasingly become an expensive game to get into as more venture capital flows to established players with traction.

-          Vaporware isn’t going to cut it like it did before.  Consumers will become increasingly satisfied with products from established (and delivering) companies.

-          I predict that large players will build screen-based devices (watches), Mid-Range players will stick to the wristband/armband, smaller players will bring devices to market with varying form factors and configurations to satisfy specific use cases.

Stay tuned for Wearable Prediction Axis 2:  Single vs. Multi-Function Wearable Devices

Cheers,

Ash

[kiwiwearables-blog1]

How to go from data to apps

One of the trickiest things in sensor based applications is identifying the trigger for your application. There is a lot of noise that exists to building application and we aiming at simplifying this method.

Currently we are looking at how to incorporate a large number of industry standard method calls for data including: orientation (yaw, pitch, roll), quaternion (3 and 6 axis), height change, velocity and much more. 

We are also looking at standard HID inputs including, left, right, up, down, forward and backward, however use this method as a single point to your body.

One of the first integrations we have built is with the visual programming language of processingjs. The first application we made was to control a dot as it moved around a screen based on accelerometer data.

The next was a cursor trail visualization using gyroscope data. 

Processingjs is a great way to get started with sensor data into a real-world feedback; check out our github page for some sample applications.

[kiwiwearables-blog1]

Using a Kiwi Move in your dev. projects, quick starting at a hackathons!

Over the past few months we have participated in a number of events whereby we asked teams to use the Kiwi move developer kit as part of their project. Some of the ideas that have come about have been fascinating, including: a behavioural change app to reduce smoking (Breath.io)

An embedded solution for pill boxes, whereby the team members quickly utilized the Kiwi move in their system as a pin out and communication channel. One team member Valeri noted: The device actually CONSISTENTLY worked it showed that it is a good product. With our limited knowledge of the device we were able to integrate it into our device in under 4 hours. Especially considering the fact that the other available of the shelf solutions failed. This is a true testament of the quality of the product. (LifeCapsule)

A team utilized it in a luggage system whereby drops, flips and orientation were detected, the quick start guide allowed the team to focus on their business case over the technology itself! (ShipMetrix)

These are just some of integrating projects that have popped up so far we are looking forward to what other ideas come about. Thanks to all teams whom have used the kiwi move thus far. Most importantly we want to thank OCAD, Startup Weekend and Mars for including the Kiwi team at their events!

[kiwiwearables-blog1]

Making "sense" of sensors

Quite a bit of noise and information is thrown at us each day and now with embedded electronic sensors this information amount is rapidly increasing. Being able to see in real-time is the first step to understanding, as sensor outputs are fed into a system a visual interface is best suited to see how movements or signals change as you alter states. 

We utilize as simple web streaming dashboard to do just that:

From the above we can identify specific motion events, for example a swinging motion. Once this "event" is identified we can take the data from that segment of time and turn it into meaningful information such as swing length and top acceleration. 

One of the trickier challenges is identifying a specific feature or event from the sensor data to allow for the event to be captured. Some of these features include: mean, average, standard deviation, covariance, and variance.

Now since each person is slightly different these features are custom to everyone so being able to find the best feature for each person to capture an event becomes even more convoluted.

Visualization and simple valuable metrics that are common place already hopefully will help people use sensors to their advantage.

Until next time: JD

[kiwiwearables-blog1]

Sensor as an advantage - sports, training

As the variety of sensors come out to the market we ask ourselves what does this data in-fact mean and how can it best be used. This is the purpose of the variety of wearable technology companies, some focusing on your sleep habits through determining shakes and orientation while you are sleeping, to others where pertinent metrics are derived such as steps or calorie tracking.

Some of the more specific metrics that have popped up recently included:

- power, velocity and force (sports and training)

In soccer/football for example specific motions events could be recognized such as: 

- kick, pass, or volley

Taking these events into account with metrics such as power and velocity players can derive pertinent information to help them improve.

Obtaining consistent reliable data is difficult task, as the placement and orientation of device continuously shifts. Any displacement in a device will yield erroneous calculations. In order to counter act this a snug holder and reliable casing is needed.

We tried our hand at capturing some data from a local soccer team. 

As we learn more about how data is captured in different worn locations we are able to help individuals learn more about their motion. 

Until next time: JD

[kiwiwearables-blog1]

Predictive Systems in a Web World

In order to properly determine a predictive system two items are needed a clean reliable signal and a strong processing platform to provide consistent and fast calculations.

For Kiwi we are still building upon both.

Here is a flow diagram outlining some of our practicing methodologies.

Until next time; John David

[kiwiwearables-blog1]

Machine Learning In the Cloud

Creative a predictive analytics engine is not for the faint of heart; thankfully the process of predictive systems with sensors has been around since the early 1900's. One of the first experiments in the realm was done by a Russian Naval officer whom wanted to predict the next best step for providing feedback for the helmsman.

The process to which we speak of is PID controllers, what these are is an iterative control feed back loop which allows for information to be continuously passed into a controller which has an optimal set point, for example a desired heading for a ship. Let's say a naval ship is desired to head true north, and in order to achieve this it sets its direction to 10 degrees off to due north to account for the current. Each time a position is taken the input to the effect of what the direction is know is feed into the system and allows for the next optimal step to be predicted.

These same principles can be applied to biofeedback systems. The underlying mathematics has been proven in many large scale manufacturing processes and has been tried & tested over the past century. Now with the advancement of reduced cost servers and extremely fast processors this technology can be made readily available to the end consumer.

The trick lies in finding the optimal set point for each individual person (a personal signature as you will) this is not an easy task to overcome, but something we at kiwi wearables hope to move the needle on little by little.

Here is a link to a wikipedia article outlining PID control systems, good read for the engineers out there. 

Next time I will provide some insightful diagrams and example javascript code of how this can be achieved!

Cheers and happy thanksgiving.

John David