/
17-6 Conclusions and Future Work

17-6 Conclusions and Future Work

Overview

This project successfully demonstrated that a Jansen eight-bar linkage could be driven by a two-motor tank-drive system to produce a stable, navigable walking gait. We used independent motor control on each side to steer the robot, and a wireless FlySky RC system to operate it remotely. Kinematic analysis was conducted to evaluate how the foot trajectory of the Jansen linkage behaved across a full crank rotation. Unlike many walking robots that rely on complex control systems and multiple actuators, our design achieves locomotion purely through mechanism geometry, with a single motor driving each side.

Our main goal was to prove that a Jansen linkage could produce repeatable, stable walking motion while still being steerable and RC controlled. We were able to successfully verify that the mechanism walked as intended, turned using differential drive, and held up through repeated testing. Overall, the Jansen linkage and the drive system functioned together as intended, and Linkage Larry met all of its primary objectives.

Lessons Learned

  1. Time Management in Design and Manufacturing Phases: This project really demonstrated the importance of allocating enough time to manufacturing and not just the initial design. While our CAD came together well, manufacturing constraints like tolerancing and assembly clearances can impact the design if they are not thought about earlier in the process. For example, we spent more time than expected getting our linkage joints to the right tightness. We learned that things like this come up quickly once you start building, so giving yourself more time for manufacturing allows for adjustments and verification before you are deep into final assembly.

  2. Getting the Kinematic Analysis and Code Right: This project also highlighted how much time and care goes into translating a mechanism from CAD into a working simulation. Writing the kinematic analysis code required us to carefully implement a dyad-based solver and correctly chain position, velocity, and acceleration through all five loops of the Jansen linkage before the output plots actually matched the expected foot trajectory. We quickly learned that getting the math right on paper is one thing, but getting the code to produce the correct motion output is a separate challenge that takes real iteration and debugging time.

  3. Selecting the Right Components: We also learned how important it is to select hardware that actually supports the quality of motion you need. Our bolt and washer pivot joints worked, but they introduced friction and required careful torque management at every joint. Too tight and the links would bind, too loose and you would get unwanted play in the mechanism. Bearings would have been a better solution since they provide consistent, low friction motion without needing to find that torque sweet spot at every joint. We would prioritize that swap in a future iteration.

Future Work Ideas

  1. Improve the Enclosure and Battery Mounting: The current design has the body cover and LiPo battery zip tied down because the cover came out slightly undersized and could not house all of our electronics. In the future, we would redesign the cover with accurate internal clearances that account for the actual size of all wired components and the battery. Adding integrated mounting features like clips or recessed pockets would eliminate the need for zip ties and produce a much cleaner final assembly.

  2. Integrate Bearings and Commercially Manufactured Gears: In the future, we could replace the bolt and washer pivot joints with bearings. Integrating bearings would allow for smoother motion of the Jansen legs and eliminate the need for careful torque management at every joint, since bearings are already pre-lubricated and provide consistent low friction motion. We could account for the additional weight by distributing it evenly across the chassis and sourcing smaller, lightweight bearings. We could also order commercially manufactured gears instead of relying on our 3D printed ones, since manufactured gears are more precise, leading to better tooth meshing, reduced friction, and improved efficiency of the drive system overall.

  3. Expand the Electronics and Autonomy: The current design already uses an Arduino and RC control, but there is a lot of room to grow. Adding ultrasonic or IR distance sensors could enable basic obstacle avoidance. A gyroscope or IMU could help detect and correct drift during straight line walking, compensating for any asymmetry between the two drive sides. We did explore some of this during the project but were limited by the hardware available to us at the time, so we were not able to test it as thoroughly as we would have liked.

  4. Reinforce the Crank: Our FEA flagged the crank as the highest stress point in the system under conservative loading estimates. A future iteration could reinforce the crank with a metal insert or switch to a machined aluminum crank entirely, which would increase the load capacity of the mechanism and give more margin for carrying a payload or operating on uneven surfaces.

Tips for Future Groups

  1. Get Your Assembly Procedure Locked In Early: This project really showed us how important it is to establish your joint assembly procedure before you are deep into building. The bolt torque sweet spot is something you need to find on your first leg, not your fourth. We recommend building one leg completely, getting it moving freely and consistently, and then using that as your reference for every leg after. This saved us a lot of time diagnosing binding issues later in assembly.

  1. Print or Cut Spare Linkages Before You Need Them: We learned that small laser-cut plywood links can crack at the bolt holes, especially if you are taking things apart and reassembling multiple times during testing. Printing or cutting a few extra linkages is not that hard and will save you if something breaks during assembly or testing.

  1. Prototype Your Enclosure Around Your Real Electronics: We made the mistake of finalizing our body cover design before our electronics were fully laid out on the chassis. Real wiring harnesses, connectors, and battery packs always take up more space than they look like in CAD. We recommend physically mocking up your electronics on the chassis first and measuring the actual envelope before committing to a final enclosure design.

  1. Start the Code in Parallel With Mechanical Assembly: While the mechanical system was coming together, we could have been validating our RC receiver signal ranges and motor direction logic on a benchtop at the same time. Starting the code early means that by the time the robot is fully assembled, the controls are already working and you can go straight to walking tests instead of debugging software at the end.

Acknowledgements

This project was challenging but incredibly rewarding. We would like to thank Dr. Meredith Symmank, Min Geun Park, David Gutierrez Moreno, and the Texas InventionWorks Team for their guidance, feedback, and access to fabrication tools. Their support in 3D printing, laser cutting, electronics, and prototyping was invaluable to the success of this project. specifically, special thanks to the TIW team for lending us an Arduino Uno after our original team-owned unit was burnt out during testing, and to David for facilitating our purchase order for the two DC geared motors and for his consistent support during our Friday morning check-in meetings.