When Dartmouth engineers designed their popular RC tackling dummy, "iterative prototyping" was the name of the game just as much as "football" was.

Over the last few years, concussions in football have become a hot-button issue. As a result, health advocates and engineers alike are looking to use technology to protect players from head injuries. For example, Riddell has incorporated sensors into helmets via their InSite system which alerts coaches wirelessly when a player experiences an unusually forceful hit to the head. However, such technologies are not a subject without controversy as the NFL repeatedly questions the reliability of such technology. 

Dartmouth football coach, Buddy Teevens, had another idea of how to address the issue. His vision was to allow for rigorous practice tackles without risking unnecessary player injuries in the process. He requested help from the Thayer School of Engineering and two students stepped up: Quinn Connell and Elliot Kastner.

The resulting product, the Mobile Virtual Player (or "MVP" for short) is a human-sized, self-correcting, and reactive RC device that can be knocked down by players repeatedly.

But getting to the advanced design available to teams today required some imagination, some elbow grease, and a lot of iterative prototyping. Here's a rundown of some of the basic challenges the team encountered in the early days of the MVP program in the labs at Dartmouth.


The Ballbot Base (AKA "What Didn't Work")

The purpose of the project was to design an omnidirectional robot which could mimic a human’s unpredictable movements. Since the robot was required to stay upright, the team thought that a ballbot would be suitable for this application. Ballbots can balance themselves on a single spherical wheel. The basic idea of keeping these robots upright is similar to that of an inverted pendulum.


A ballbot uses a feedback mechanism to remain upright. Image courtesy of robohub.


Although these robots are omnidirectional, agile, and maneuverable, there were certain challenges in using a ballbot as a robotic tackling dummy. Firstly, they're actively trying to stay upright. This is not a desired feature for a tackling dummy because players are not supposed to fight with a robot which is dynamically stable.

Secondly, they needed a robot which could pop up immediately after each tackle. And thirdly, after experimenting with many different air-filled balls and 3D-printed components, the team concluded that a ball drive led to a lousy traction on grass and tended to drift rather than turn sharply.

So, according to Quinn Connell, now MVP's Director of Engineering, "We initially started with a ball drive, but the challenges of the interfaces led us to revert to a more traditional wheeled design. We found ourselves at the limits of the electric motors that are available and ended up custom designing our own motor and drive train to fit our specifications."

This change came at the cost of being able to utilize a truly omnidirectional system but allowed for much more realistic movement for the dummy overall.


Adventures in Prototyping: Making Old Ideas New Again

The team employed a passive mechanism to keep their robot upright. Interestingly, the stabilization mechanism they were looking for has been used in several lines of children’s roly-poly toys with the trademark of Weebles, which was first introduced in 1971. "Weebles wobble but they don’t fall down” was the popular catchphrase of these toys during 1970s.

A Weeble has usually an egg-like shape with an almost smooth bottom hemisphere. As shown in the following image, the lower part of a Weeble is made of a material (m2) which is much heavier than the material used for the upper part (m1). This puts the Weeble's center of mass very low. In this way, when the Weeble is "wobbled", the center of mass exerts a force and the hemisphere-shaped bottom allows the toy to roll back to its stable position.


The operation principle of a Weeble. Image courtesy of K.D. Schroeder [CC BY-SA 3.0], via Wikimedia Commons​


The tackling dummy started out based on this design, using available off-the-shelf components. Pouring green foam into a mold, the team built a human-height cone with a rounded base which used big chunks of steel to stabilize it.

They also originally used an off-the-shelf radio communications system to control the robot remotely and used a car battery to power the heavy robot and run the initial tests. But the robot was (unsurprisingly) not agile enough and there was still a long way to go before they could claim to have a truly functioning prototype.


From Academics to the Market

Launching a Kickstarter campaign, they raised $5,000 to further develop their device. The video on the page attracted a great deal of attention and many communities and coaches clamored to get an "MVP" dummy for their team.

The prototype they released was about 70 kilograms and had a top speed of 32 kilometers per hour, which was close to that of a human player. It could rotate in place and weave between obstacles as directed by an RC controller on the sidelines.

For more on the development of the MVP device, including how they handled crowdfunding and public engagement, don't miss our full interview with MVP's Director of Engineering, Quinn Connell, next week on AAC!