MKS INSTRUMENTS MECHANICAL/ELECTRICAL ENGINEERING INTERNSHIP.
3D Modeling
The final deliverables of the Smart Bike Light Project involved a video demonstrating and explaining the inner workings of the Arduino light. However, the manufacturing of the metal aluminum outer housing was difficult to machine using the mill which made fitting around the 3D printed difficult. Thus, my solution to fast track the video production was to animate our product. This came with a few advantages which included showing the audience the inner sensors and controller boards to better visualise how the light worked as well as allowing the team to physically machine, assemble, and test the bike light while the video production was underway in parallel. Only then would the physical bike light be shown operating in the video along with a live demonstration during presentations.
​
Since CAD models were already made of the bike light, I used Blender which allowed for STL import file types. This would allow myself to work on texturing, UV wrapping, and key-framing without remodelling each part again in Blender. The result produced fairly accurate models with aliasing along rounded corners, a byproduct of the STL format.
SOLIDWORKS STL File

Blender Import
UV Mapping and Texturing

The imported parts from the CAD files needed to be textured and coloured for the final render. For some of the parts, including the 3D printed inner structure and battery casing as well as the aluminum tube, the material properties were replicated using shader nodes to define the shine of each part. For example, the aluminum casing is fairly reflective with some buffering on its surface, thus a low anisotropic value of around 0.4 with a roughness of around 0.3 was used to achieve the final texture.
​
For the batteries and the LED driver (seen above), labels and stickers had to be replicated to represent the real-life parts. Thus UV wrapping was used where the images of the labels were imported and mapped onto the surfaces of each part. As a result, the parts better represented the physical counterparts however, specular detail, specifically with the solder points of the Arduino Nano were difficult to properly replicate due to the granular STL tessellation. For the sensors of the GPS Module and Light Sensor, no CAD model existed, thus a basic model was created under the timeframe of the project.
​
If I were able to redo the texturing of the parts, a full remodelling of each part would reduce the aliasing as well as UV wrapping inconsistencies. Also, developing a background environment to better showcase the product would add better detail into the final video.
Keyframing

Keyframing and timeline management added complexity to the task of animating the video. Thus, movement of each part was reduced to no acceleration. This made the animation robotic rather than smooth for not just the parts but also the camera. However, under the time constraint of the project, this was a necessary compromise.
​
Relative movement of the parts was founded to be quite difficult, especially with the lower clamping mechanism. This was where a metal pin (seen above) would fit through a hole allowing the clamp to open and close. When the clamp was in focus of this part of the bike light, the clamp assembly was rotated in view of the camera's perspective. This meant the opening and closing of the clamp had to be animated at an angle which proved difficult since the axis of rotation for the bottom clamp was located in the centre rather than at the pivot point with the pin. As a result, the lower clamp can be seen clipping through the upper clamp.
​
If I were to redo the animation, I would implement acceleration for smoother movement as well as learn how to restrict parts to relative motion to prevent clipping.
Rendering
Finally, the Blender animation was rendered in 1080p using the H.264 codec with the product being compiled using a basic editing software along with extra filmed clips seen at the end. Recorded audio was then introduced and was uploaded to Youtube for the final presentation of the project.
​
All in all, I am proud for my ability to learn, texture, animate, script, record, edit, and render using Blender all in under 6 hours. My solution of animating portions of the video to better showcase how our team's product functioned was adopted by only our team and was praised for its creativity.

