During the prototype phase, our team has gone through multiple iterations for our design. Our process began with the initial brainstorming found in the project proposal which used the analysis of the following link lengths (L1 = 32mm, L2 = 16mm, L3 = 32mm, L4 = 20mm, LP (ternary) = 40mm) to determine the geometry of each arm that the climbing robot would have. Our initial design decided on having 6 legs that would have different instances of rotation to allow for constant contact with the wall, however, we decided that doubling up and making two layers of legs (12 legs total) would allow for a more even distribution of force and fully ensure that there would always be contact with the two parallel walls. To allow for variability in the leg's distance from the wall, a spring and rubber foot will be implemented at the end of each leg for compliance. This will keep the system from receiving unnecessary stress and be able to keep each individual leg on the wall for a longer period of time.
For powering all of the individual legs, it was decided that using a gear train would be the easiest and most efficient to implement. This would allow for all twelve legs to be powered by one individual motor. The gear would act as link 2 with there being an offset location from the center for the ternary link (L3) to connect to. Figures 1, 2, and 3 show initial sketches from team members regarding how the legs and gears would be implemented.
Figure 1: Gear system and leg link order
Figure 2: Gear system with spring and rubber feet
Taking these dimensions and ideas, the first prototype that was created was a single leg to ensure that the motion was achievable. A CAD model was designed for just one leg and includes an input gear, an idler gear that transfers motion from the input, and a leg gear (L2) that receives the transferred motion to move the leg. Figure 4 below shows the initial model in Solidworks.
Figure 4: Single leg prototype
The model was then created by laser cutting all of the parts from scrap wood at TIW and assembled using screws and wooden dowels as axles. The result can be seen in Figure 5 below. The result was a design that was not very stable and often had gears come loose. It was determined that bearings were needed to get better motion and ensure that gears and links stayed in place. The prototype however was able to achieve the motion determined in the initial analysis.
Figure 5: Assembled single leg prototype
With further analysis, it was determined that improved motion could be found by once again changing the link lengths to get a longer and flatter vertical movement for each leg. The new link lengths that the second prototype used were (L1 = 30mm, L2 = 10mm, L3 = 30mm, L4 = 12mm, LP(ternary) = 65mm, theta between LP and L3 = 80 degrees ). This greatly improved the motion as can be seen in figure 6 below. This prototype has 4 legs powered by a hand crank that simulates the input power from a motor that the final iteration will use.
Figure 6: Four leg prototype
The Solidworks assembly was once again laser cut and assembled but just with two legs to ensure that motion on the input side could be translated to the opposite side properly. Figure 7 below shows the completed prototype in motion. This iteration uses bearings to smooth the motion. The design also uses wooden dowels as axles that are press fit throughout the prototype to keep everything well put together. This final prototype iteration should prove that it is possible to do all of the required 12 legs. For the final product, springs will be added to the ends of each leg with rubber tops to improve compliance with the parallel walls that the robot is climbing up.
Figure 7: Assembled two leg prototype.