Our Design Process (and How it was Used to Design Gnorp)

By Lucy C.

Picture this: a group of enthusiastic grade school students decide to create their own robotics team. This team has all the tools and skills it needs: sufficient hardware components, apt software, proficient analytical abilities, and an inclusive, STEM-oriented environment. What could possibly go wrong? Everything. Every week, the students gather in a frenzied, disorganized mass. They scramble to meet deadlines and jump from one task onto the next in a flustered, uncoordinated manner. They could barely make it to competitions, let alone in Regionals! By the end of the school year, all that is left a disbanded team and a bunch of befuddled students.

Chances are, these students did not have a structured, well-defined design process to follow. While any FTC team may harbor bountiful skills and resources, these skills and resources will not amount to anything unless the team has a strategy to put them to good use. This was why these students struggled to maintain their team: they could not follow an organized process or workflow that allowed them to actually utilize the abundant skills and resources they accumulated. This is why CAOS Robotics has learned to establish a clear, structured design process to better optimize our own skills and resources.

This design process is a six-step process called the Design Hexagon. These six steps comprise a sequential yet cyclical process that ultimately revolve around the improvement of the robot:

Here is how our process works:

1. Define design criteria.

   In this step of the process, we define how we will design our robot. We take note of game parameters we must fulfill for the season, the hardware and financial resources we will need for the robot, and the amount of time it will take to design each component of our robot.

   
   

   For the design of Gnorp, for instance, we took note of that season's STEAM-themed game parameters and developed a spreadsheet that listed basic materials and costs needed for us to successfully fulfill these criteria.

1. Research game parameters and past robots.

   Here, we must further inspect the season's game manual, particularly for the scoring and mechanics of the game. That way, we can research past robots that for software or structural inspiration specifically tailored for optimizing scores. We may thus extend our research to the various scientific and mathematical principles that may be relevant for incorporating these sources of inspiration into our actual design.

   
   

   While conducting research for the design of Gnorp, we recalled how code for the autonomous periods of previous robots like the Picky Eater utilized the motion planning library Road Runner. However, after reading the corresponding game manual and learning about the intricate layout of the field, we decided we needed greater control and flexibility over Gnorp's autonomous movement. This was why we researched mathematical principles like Bezier splines in order to better customize the robot's pathfinding and object avoidance algorithm.

   
   

2. Brainstorm potential design choices

   We dedicate multiple days---often weeks---to freely sharing and presenting design ideas, regardless of how ridiculous or ill-conceived they may sound at first. In fact, we brainstormed three different designs for Gnorp's outtake: 1) a claw that would be positioned vertically from a single pixel, 2) a claw that would be positioned perpendicular to a stack of pixels, and 3) our current design, a dual outtake that could insert itself into the centers of up to two pixels.

   
   

   1) Vertical Claw

   

   ⬇️

   2) Stack claw

   

    3) Dual insert outtake

   

   
   

3. Conceptualize the design

   During this stage in our design process, we use computer-aided design to conceptualize the design of our robot and tailor our code to this design. That way, we can determine whether the ideas we have brainstormed in the previous step are architecturally possible and meet our design constraints. Before designing our robot, we used Onshape to model its entire assembly:

   

   Onshape model of Gnorp

   
   

4. Build the robot! 24+ hours are dedicated to the actual construction of the robot. We must apply necessary engineering and mathematical principles we researched in Step 2 and the design choices we conceptualized in Step 4:

   

An image of our actual robot.

1. Testing and Tuning

   This is where we test our robot to see if it can execute necessary functions. We may continue "tuning" and refining our code to better align this design. In case we test our design and find it unsatisfactory, we can revert to previous steps of the Design Hexagon so that we may adjust our design accordingly. This step thus contributes to the cyclical nature of our design process and allows us to continue improving and upgrading our robot.

   While designing Gnorp, we constantly tested and refined our robot. As previously mentioned, we initially designed and tested a vertical claw mechanism for our outtake, only to revert back to Step 1 and conceptualize two other outtakes that would better help us meet game criteria. Our robot initially also had a beeline drivetrain, but upon testing this drivetrain, we decided it was too heavy, bulky, and inefficient. Hence, we instead designed a mecanum drivetrain, which had a compact, more streamlined design that allowed greater precision in movement.

   

   To the left is our robot designed with the Beeline kit. To the right is the Mecanum robot. We preferred the Mecanum robot for its compact design and advanced feautres.

An effective design process is arguably one of the most important aspects of being successful in FTC. As a member of the FIRST community, CAOS Robotics knows that the raw accumulation of knowledge and resources is not enough to thrive in all of its Tech Challenge. Rather, we must be able to strategically manage and utilize our knowledge, time, and resources with the help of a design process. This is why we use the Design Hexagon.