Home / Past and Future RoBOPPics

The granting of the 2013 USPTA Patent US-20140038751-A1 Personal Sports Simulation Robot. is the seminal event that began the development of the BallBOPPer Autonomous Tennis Training Robot. 

NOTE: The BallBOPPer started out in 2013 as a part-time side project. Full-time work on the BallBOPPer robot began at the beginning of 2020.

The following drawing from the original patent looks quite different from our current prototype, but the underlying design and function are still fundamentally the same.

The invention as described is "a personal sports simulation robot that is able to simulate the competitive play of a human opponent with particular playing characteristics such as a particular playing strategy, a particular physicality, a particular playing style, and a particular skill level throughout the course of a simulated game or drill."

Today we simply call this True Game Play which is envisioned as the ability to take the robot out to any tennis court, turn it on, and be able to play tennis: points, games, sets and matches.

During the early years many different prototypes were developed and tested that explored alternative form factors. A principal design goal for the robot is to be "personal": small, light and easy to transport to and from any standard tennis court. The following three prototypes are representative of the many different prototypes that explored different ways to accomplish this goal.

The prototype on the left most resembles the ball launcher and bucket shown in the patent. The center prototype most resembles the multi-motor, self-steering drive system shown in the patent. The right prototype was one of several different "luggage style" prototypes (pictured without the fabric outer). There was also one fully functional prototype, not pictured, that was built into an airline carry-on case.

In the end, all of these conceptual designs proved less than ideal. Too small or too large and heavy, unable to satisfy the requirements of a fully autonomous mobile tennis training robot. 

A series of prototypes followed that developed more organically by letting "form follow function". Each of these prototypes were aimed at the perfection of the specific set of capabilities required to achieve the BallBOPPer's design goals. 

This approach led to the prototype shown above which was focused on the perfection of the compact and modular ball launching head, and the perfection of the removable, ball-hopper-like ball bucket and ball feeding system.

This prototype wasn't mobile, it wasn't smart, it only launched balls up to 3/4 court, but it was a quite nice and serviceable little ball machine.

These two capabilities, the easy to reload ball bucket, and the compact, easy pan and tilt launching head, formed the basis of the first fully computerized robotic systems developed between 2020 and 2023.

These prototypes developed the roving system, the reloading system, and the Vision AI localization system, making the BallBOPPer fully autonomous for both the launching and the reloading of balls.

With the 2023 prototype the system also became completely modular. 

The ball launcher and rover as a single unit fell short of the "personal" design goal. The combined size and weight of the rover, reloader and launcher were enough to make the unit challenging to transport to and from the court. This was resolved by making them separate units which snap easily together on the court. 

This change not only restored portability, but it has the added advantage of allowing the Player to use the launcher by itself for skills training when the roving and reloading capabilities do not necessarily add additional value. 

Underneath the covers of all this is of course the software, both conventional code and machine learning AI models. These play an enormous role in the BallBOPPer's ability to deliver on it's design goals.

Without all of the recent developments in Machine Learning, and Vision AI in particular, there isn't any way the BallBOPPer and True Game Play would be remotely possible. The BallBOPPer product is truly riding the leading edge of the "great wave of technological innovation" that is being driven in large part by Artificial Intelligence and Autonomous Systems.

The smartphone app, a few screens of which are shown above and which has been out in beta on Android for several years now, is perhaps the most sophisticated tennis training app in existence. The capability of configuring Patterns of Play (PoP), including the innovative features Repeating Ranges, Alternating Shots and Voice Control, has proven both reliable and easy to use.

But even with all of these accomplishments and groundbreaking capabilities, the BallBOPPer has not yet completely delivered on True Game Play.

Which brings us to the Production Prototype which is currently going through our rigorous testing regimen. The Production Prototype has two new capabilities which we have yet to reveal. These two capabilities will give us the complete system needed to deliver True Game Play. The BallBOPPer will then only require that the Player/Sponsors for the crowdfunding campaign actively use their BallBOPPer systems for on-court training and generate and contribute their on-court playing data. As more data is gathered it will be used to train the TGP model and True Game Play will then light up on its own.

And what comes after the final delivery of True Game Play?

Strategic Autonomous Play (SAP) is in the planning stages. This is a computational decision-making model that is quite different from Patterns of Play. When you decide to play against the SAP model, the BallBOPPer will play at your skill level, but the SAP computational model will make tactically and strategically optimal decisions. In other words, the skill level will be identical to yours, but the tactical and strategic side of the game will challenge you in a completely new way. We expect to learn great things from it. It will be the first AI model to bring strategic computational "thinking" along the lines of Deep Blue to the playing of a physical sport.

Looking even further down the line, we continue exploring and researching a potential combination with Augmented Reality (AR) technology.

VR, AR and MR headsets are currently complicated and cumbersome, but there is a tremendous amount of ongoing research and development that is shrinking their size while simultaneously increasing their capabilities. Within five years there will be devices with the level of performance required, in the form factor of a regular pair of glasses that will be practical to wear while playing tennis and other sports.

There are affordable consumer AR headsets already available with the eyeglass form factor - such as Meta Ray-Ban, Lenovo ThinkReality, XREAL Air, RayNeo X2 etc. These are capable of overlaying digital content onto your view of the world, but they fall short in terms of compute power, field of view, interactive capabilities and suitability for sports activities - all of which the BallBOPPer AR integration will require. 

As soon as affordable fully sports capable AR glasses arrive, RoBOPPic's is planning to deliver application software that can superimpose an animated generic or specific player experience such that when you are using the BallBOPPer, instead of seeing the BallBOPPer at the other end of the court, you will see the player you are training against moving around the court and executing each shot wielding a regular tennis racket with realistic tennis bio-mechanics. 

To be sure, the proposed integration of the BallBOPPer with AR is non-trivial, but there are a number of companies making substantial investments in VR, AR, and MR technologies, and it is inevitable that this capability will become viable before too long.

True Game Play is our primary focus right now, but we believe that by 2030 we will be able to combine TGP, POP and SAP with AR to deliver an exceptionally realistic, fully physical, on court tennis training experience.

******

To receive an invitation to the Crowdfunding Campaign, please subscribe below.

You can also follow or link with RoBOPPics on Facebook or LinkedIn.

If you have a specific question, you can send an email to info@roboppics.com.