Please note as of Wednesday, August 15th, 2018 this wiki has been set to read only. If you are a TI Employee and require Edit ability please contact x0211426 from the company directory.

Chronos High 5 Me!

From Texas Instruments Wiki
Jump to: navigation, search
The #evp parser function was deprecated in EmbedVideo 2.0. Please convert your parser function tag to #ev.
 TI joins the social network revolution as "Let's High Five" integrates its cutting edge technology in a social application to celebrate happiness. Using the Ez430-Chronos watch, anyone can reach different levels of popularity and meet new people through high 5s. Our pattern recognition algorithm enables this watch to detect high 5s and wirelessly syncs with your computer to automatically share your accomplishments with your friends and the rest of the world. Join our community, boldly celebrate your happiness, and bring the high 5 back.

Follow us on twitter: TI_HighFive



  • High 5 Detection and Counter on Chronos Watch
  • Automatic posting to Twitter and letshigh5 website
  • Website Leader Boards
  • Social Network integration

Let's High 5 Social Network

The Web 2.0 and social networks have revolutionized how we communicate, interact with each other, and take on our daily lives.  With more than 500 million users on Facebook, there is no doubt that the market opportunity in social media / applications is huge.  The emergence of the Iphone, smartphones, and other internet enabled phones have created on us a psycological attraction to express through blogs, tweets, etc.  There are networks where people can post their current location, pictures, their current status, etc. 

Wearable electronics bring forth innumerous applications and potential for social media.  Let's High 5 takes advantage of ez430-Chronos' built in sensors and features to provide a unique application that cannot be implemented using a traditional iPhone or smartphone.  By using the acceleration data in the wrist of a person, our application can detect its movement and decipher whether the person has just high 5ed. 

The value our application brings to Texas Instruments is remarkable.  We want to remove the misconception that Texas Instruments = Calculators.  For the least minimum, we would like to be seen in public light as Texas Instruments = Calculators + Super Awesome Watch.  With Let's High 5, TI can expose its technology through social media and a fun popular application suitable for anyone.  Wouldn't TI want to be the reason for the high 5 to regain popularity?

Our website features a simple style, with top 10 high five leader board.  Here, people can expose their awesomeness and popularity from the amount of high 5s they have received.  Even more interesting is the automatic integration of the watch and Twitter, where your current high 5s will be tweet and exposed to the world.  Nevertheless, the potential of our social application is endless. 

Imagine logging in to your favorite social network and seeing your news feed filled with your closest friends recent high 5s.  Imagine meeting people simply through high 5s.  Instead of getting their name, or looking them through facebook, a high 5 between two strangers with the chronos could automatically trigger an exchange of social profiles.  This way after you high 5 a stranger in a bar you'll get a friend request and you have expanded your social network.  Imagine combining geolocation data of every high five and being able to map geographically every high 5 in realtime.  This way you can see where your friends are having the most fun, or what bar currently has the most happy people.  With the ez430-Chronos, this and much more is possible.

LetsHigh5.PNG             Twitter.PNG

Visit us, leave us your comments on how to improve:                                Follow us on twitter:

High Five Detection Algorithm

Several methods for the detection of high fives were investigated throughout the course of this project.  One of the first methods we tried was to look at the overall magnitude of the acceleration.  If the magnitude of the acceleration exceeded a certain threshold the program interpreted the event as a high five.  While this approach was computationally simply and provided reasonable accurate high five detection, it was difficult to reject other events such as clapping and and shaking.  

Next we investigated using the altimeter in an attempt to detect when your hand was elevated.  However the low sample rate of the altimeter ment that it was not suited to this application.  In addition an elevated hand was typically only one code different from a low hand.

Finally we began to investigate correlation based high five detection.  In this approach the unknown event is compared to a known event.  If the two events are similar the correlation between them will be large.  With this approach an event can have a high magnitude but still have a low correlation.   To generate the know events we sampled different high fives and averaged the events for each axis where x and y are the sample means of X and Y, sx and sy are the sample standard deviations of X and Y


The data was processed and plotted using MATLab.  The following screenshots depict the mean behavior of the high 5 in each of the different axis.  A trend is observed and precisely that behavior is being used to filter out any non-high5.


                                                                       Fig. 1: Matlab plots for the average of a high 5 in each axis

After the correlation algorithm was established and tested, different data points were used to establish a threshold for a high 5.  This filters out any behavior in each axis different from the ones depicted in the figure above.  Before implementing the algorithm in the ez430-Chronos, a program in LabView was written to test and visualize its performance in real time.  The following screenshots show the main features of the circuit block.  The source code is found below.

                Labview1.png                                                            Labview2.png

              Fig. 2: First three plots are realtime accelerometer data, bottom plot is result of correlation.                          Fig. 3:  High 5 behavior and the coefficients for each axis

Usage Instructions

Labview Code

Front Panel.jpg
Installation Instructions

  1. Download and install the Labview Evaluation Software
  2. Download and install NI-VISA
  3. Download and install the drivers for the USB RF module
  4. Download the Labview Source Code
  5. Open the device manager and locate the COM# for the TI C1111 USB RF module, type this into the Serial Port control located on the front pannel of the Hi 5

How to Collect Acceleration Data

To collect acceleration data from the watch first open Hi 5 and click the run arrow in the upper left corner of the front panel.  If a watch is within range the connected light will light up.  Next go to the ACC menu in the watch and press the lower right button to begin sending data.  Click the Start ACC Data Collection button to begin collecting data from the watch.  The first 3 graphs display the X, Y and Z acceleration data. When the value in the correlation graph exceeds the detection threshold the High Five! indicator will light up.  To stop collecting data simply press Stop ACC Data Collection button.

Python GUI

Installation Instructions

  1. Download and install python from (Link)
  2. Download Python Source from below


Install, setup RF access point.

Press connect.

Connect watch.

Programing the Watch

Installation Instructions


Go to the Hi 5 mode on the watch by repeatedly pressing the lower left button.

Start high 5 mode by pressing the right lower button. (a heart and your total high 5 count should display).

Once you want to send your high 5s to the PC, click the right lower button again.

Future Work

Improve the High Five detection by having correlation functions for other events that are similar to high fives such as punches and clapping.  The algorithm can compare the correlations of different events and then select the highest correlation.   This way we can provide a better detection algorithm while filtering other types of movements.

Source Code