Fencing Strip Build + Robot Design

Fencing Strip + Robot Design

Summer 2020

Motivation

I really enjoy fencing, and I have been doing it for 7+ years as of 2023. Over the pandemic it was difficult to practice due to the lockdown, since fencing is inherently a two person sport. While I did not really have the means of building the robot, I set out to design a sort of fencing robot partner that could be used as an interactive dummy. Around the same time, a friend and I built a strip in my backyard so we would have a place to practice while the clubs were closed. I envisioned some ways that I could make a robotic opponent travel along this strip, but due to time and budget reasons I only ever got around to making the strip.

Fencing Strip

A regulation fencing strip is 14m long, with an additional run-off of 1.5m on either end. This amounts to a lot of 2x4s, plywood, and paint needed to create one, especially one that is elevated on uneven ground. The strip was constructed to use 4ft x 8ft sheets of plywood as the top surface, since these were readily available at my local Home Depot. We chose an 18m total length which could be made with 7 full sheets and a small extension at either end. Under each sheet was 8ft long 2x4 placed in 1ft increments as if we were framing a wall. Smaller 1ft sections were added in between, and either end had a 4ft section across it.

We started by leveling a string across the distance that the fencing strip would span, and building the first framed section. Standoff lengths from the ground were created by measuring up to the level line in multiple places. The standoff sections were created out of pressure treated lumber, and these were attached at the intersections of the existing frame using a nail gun. This process continued down the whole length of the strip. First a frame is assembled, then it is attached to the previous section at one end and held to the level line, then the standoff distances are measured at a few locations so they can be cut and attached. The result is a full length leveled frame that is ready for the top surface.

The 2x8 sheets of 3/4 in plywood were attached to the framing with decking screws, and the physical strip was complete. To finish the project, decking paint was added to the top, then markings at the proper intervals. The underside was sealed with a water based polyurethane for protection from the elements. When not in use, a tarp made by tying together multiple sections was placed over the top and strapped down.

The entire project from start to finish took about 3 weeks, but the end result was very sturdy and perfectly usable for fencing. Overall the result was better than I imagined, especially since we did not have any experience with construction or framing when we started.

Fencing Robot

To begin the robot's design, I tried to define he range of motion required in fencing. I fence sabre, which is only one of the 3 disciplines, so I set out to make a sabre fencing robot. Sabre primarily requires a cutting or 'snapping' motion rather than a thrust, and it also involves parrying motions where the blade is perpendicular to this cutting motion. To start off I tried to satisfy the three most common cuts and parries.

My initial sketches included a pulley based shoulder, but I decided I a pneumatic shoulder would be faster in the end. I also did not draw the shoulder joint or the grounded link.

To verify the motion was as I imagined it, I tried to create a realistic CAD model of just the arm itself. We had been learning Fusion 360 in my engineering class in high school, so when I set out to make more than sketches on paper of my fencing robot design, I chose to do them in Fusion. The model is somewhat crude in places because of my limited design experience at the time.

retracted position

Fencing blade attached to the pivot

A later version of the model which was more complete

I did not plan on building this robot, and I did not even own a 3D printer at this point, so I mostly ignored manufacturability during the design process. The resulting robot was a very simple arm with 3 degrees of freedom from 2 pneumatic actuators and a geared DC motor that is sometimes used in FRC competition robots. The angle of the cut is determined by the DC motor, and the snap is actuated by the small cylinder in the wrist. There is a sort of 3 bar slider linkage in the wrist that constrains the angle of blade throughout the cut. The larger pneumatic cylinder is only used to lift the arm and parry a cut aimed at the robot dummy's head.

The 3 bar linkage is designed to replicate the way that a fencer extends their arm at the same time as they cut. Since these two motions are usually linked in fencing, it seemed reasonable to reduce complexity by linking them here.

Since I was not building the robot I decided to imagine what an ideal version would look like with an ultra-light carbon fiber slider

Shoulder with wrist motor

The arm in the head parry position

Results

While the robot was never built, it was a helpful exercise in design and rendering in Fusion 360, which became my go-to CAD program after that. The more I learn about robotics the more I realize the many ways this design could be improved, and eventually I hope to build a much better thought out version of this.

I see some interesting applications of a fencing robot from training to teleoperated bouting, and maybe even using AI to analyze and react to create the ultimate opponent.

Page updated

Google Sites

Report abuse