Brian Gordon and
The Simi Tennis Application

Sean O’Neil

Dr. Brian Gordon is taking it to the next level. As the premier quantitative researcher in tennis biomechanics, as detailed on Tennisplayer in numerous articles (Click Here), his work is not primarily academic. He uses it in building strokes as a full time working coach.

So what's the next level? It's the holy grail of 3D research: a completely markerless data capture system. Historically the process of 3D scientific measurements has required that players be fitted with reflective markers or sensors which can interfere with performance in producing representative tennis strokes.

Limitations

Physical markers required a long set up time. Wearing them could also be an unnatural and often uncomfortable experience for the player when trying to hit strokes.

In August of 2013, The New York Times wrote an article about Brian's research entitled "For Tennis Players, a Tech Lover's Paradise." The article describes how Brian had to physically place up to 70 markers on the body and racquet of a student to be able to create any data points about an individual's stroke.

With the markerless system, the player simply walks out on the court and hits normally. This allows data collection under any conditions including match play. That is a huge difference.

But This new system goes further. It combines the technological breakthrough of markerless motion capture with tennis specific biomechanical information based on Brian's decades of basic and applied research.

The technology was developed by Simi Reality Motion Systems, which is based in Germany. Brian has worked with them on the system for several years.

While there is still work to do, eventually it will be available as a turnkey tennis teaching system called "Simi Tennis" for use by any tennis researcher, coach, or academy.

As the graphic shoes, the system produces a three dimensional quantification of the stroke. It produces scientific grade data and information that can be used on the court in real time. It's the most sophisticated tool that can be used for stroke development.

Limitations of Video

Video, for example, is two dimensions and can show motions in detail when shot at high frame rates, as in the Tennisplayer archives. (Click Here.) But video can't say anything about the underlying biomechanical processes or the speeds of movement of the body and racket.

"With our video we can identify the key positions in the various strokes and use them as models," said John Yandell, founder of Tennisplayer.net which has a vast library of clips of the strokes of top pro players.

"But what video can't do is tell with certainty what is actually going on in the body in terms of muscles, bones, or the precise joint angles in all the strokes."

"Is one particular stroke pattern more efficient and effective? Do certain patterns increase the risk of injury?" Yandell asks.

These are questions that can be answered through 3D analysis. This is because quantification produces precise 3D coordinates for key anatomical landmarks on a player.

This creates the ability to conduct computations that allow assessment of the musculoskeletal processes necessary to high level stroke production. These may be basic things like velocity and acceleration (kinematics). They may also include assessment of forces and torques (kinetics).

The Simi Tennis system combined with Brian's research can produce computations that are directly relevant to stroke development by focusing on empirically based interpretations of commonly taught tennis concepts. This is a critical link between the science and the coaching that has been a roadblock to technical advancement in developing players.

The lack of hard measurements probably also contributes to the huge diversity of opinion and the disagreements among coaches about how to hit the tennis ball and what to emphasize in oncourt teaching.

Simi Tennis

Fast forward to today, with the new Simi Tennis technology no markers are placed on the player. A player hits ground strokes in normal attire Each camera automatically identifies the anatomical structure (lines on player). Software reconstructs the information from each camera to produce a 3D mapping (shown by yellow stick figure) in virtual real time.

The video and the quantification are from Brian's location in Boca Raton, FL. His tennis Center for Performance Research at the Rick Macci Tennis Center is arguably the most advanced applied research court in tennis.

The court setup is as follows: 8 cameras filming at 200 fps surround the court at a height of 13 feet. The cameras are connected to the main computer for synchronization and data transfer through a system of cables.

The primary controlling computer displays all camera views and produces the 3D reconstruction of the 2D videos in virtual real time. This allows the biomechanical computations to be made on the fly.

The 3D stick figure graphic along with access to graphs for kinematics and Brian's specialized data are immediately available. Brian can use this information on the spot to work with players.

For example, a player can execute a series of shots while being recorded. Immediately following the stroke set, the results can be discussed by analyzing the results on the computer at court side. Things like segment rotation speeds, joint angles, racquet speed sources, angular momentum flow or excessive joint loading may be addressed. Then, back to the court to correct any problems.

Motion capture assessment and intervention in sports is rare--and even more so in tennis. The new markerless technologies are in their infancy.

While the diagnostic and interventional capabilities are unmatched, there are some problems: (1) the systems are expensive (2) the systems are difficult to operate (3) the data is difficult to interpret. Together with Simi, Brian's operation in Florida is the staging ground to perfect the system and protocols for potential wider scale use.

Brian's overall vision for the system goes well beyond the implications for hands-on player development. In fact, that is secondary to the main goal of continued applied research and data-based information dissemination.

The new technology, for example, makes it possible to collect stroke data on large numbers of players, including live data from pro level events. This will allow Brian to take a deeper dive into the kinetic (force, torque, angular momentum) information inversely available through the measured kinematics (position, velocity, acceleration).

The result will be quantitative blue prints for each stroke at every developmental stage. Imagine having a library full of "models" for any age, skill level, type of swing--and being able to compare an applicable model with the video of any individual student. By combining hundreds of samples these are stroke models that can go beyond even the high speed video models on Tennisplayer.

The practical uses of this type of information are limitless and could greatly improve the ways that tennis is taught. I'm very excited to see how Brian's vision progresses in the future and the possibilities that will present.