Lieutenant Colonel Brandon Newell and I joined @chrisblower at FATHOM to see the MLV build status. The frame has come a long way since the initial CAD renderings! Check out the construction module and platform:
We wanted to use this blog entry to unpack some of the major design decisions we made that deviate from @saintarnab’s HAT 2.0. As you may recall from earlier blogs, the goal of this build is to quickly prove the concept of modularity.
At this early stage, many of the decisions were made to match the rapid pace of this build (from winning design to driveable vehicle in 10 weeks!)-- for example, we employed a 4-wheeled electric donor vehicle to drastically reduce the build time, but this also meant that this build couldn’t follow the original 3D printed space frame design with hub motors and independent variable suspension on each of the 6 wheels, at this point in the game. Similarly, a number of the design decisions below were made because we had very little room for critical errors-- everything needed to fit properly and function correctly on the first try, or be designed in such a way that it could be fixed on-the-fly.
Here are some of the major design decisions we’ve made that deviate from the original design:
Wheel flanges: We switched from 3D-printed reinforced polymer to oversized reinforced steel (tubing and plates) to ensure that they won’t crack during shipping, handling, and (loaded) module swapping. We’ve made an effort to maintain the original design because, not only is it a key feature of @saintarnab’s unique aesthetic, we have also made it functional to guide and center the modules during the module swap. We’ll also add a layer of delrin, an ultra-slippery polymer, to the top of the platform to take the compressive load of the module instead of metal-on-metal contact. (Note that the base will be powder coated black, keeping with the original color scheme.)
Actuators: We’ve purchased these actuators that hold ~2500 lbs each (you can see the top of the actuator I’m holding in the picture below, and we’ve got four in total, one for each corner), so we can be confident they won’t bend in the field. The actuators extend by a threading mechanism, so that they are more resistant to dusty environments and will hold in place even if a module swap is interrupted mid-swap. The actuators are remote controlled, so fewer Marines are needed to swap a module. Instead of a retractable system to store the actuators within the module frame, we’ve opted for a rigid attachment of the actuator directly to chassis (see the little white flange next to Brandon’s left hand) to minimize potential failure points. The omnidirectional wheels in the original proposal require a flat surface and can also be tough to maneuver with heavy modules, so we’ve swapped these out for rigid stand points-- extensive safety testing and stress analysis would be required to enable the installation of uni- or omni directional wheels to be installed prior in demonstrations or employed in day-to-day operations..
Windshield: The original design called for a bent windshield to extend the field of view to include above the module, which we liked because it allows the Marines to see more as they navigate around base. However, a 90 degree bend radius would require us to compromise on materials-- for example, acrylic is bendable but damages easily and would probably “spiderweb” at that bend radius, whereas polycarbonate that’s thick enough for environments like those the MLV will be used in (0.5" thick) won’t bend easily to that degree. Instead, we’ve opted for polycarbonate in two pieces that join at 90 degrees (at the horizontal bar just above my right hand). This also means that we can move the hinge on the trunk to that same edge, to reduce the leverage it experiences and thus also the likelihood of failure.
Driving mode: Arnab’s original design ran in autonomous, semi-autonomous, or wireless driving modes (using live video from cameras on the frame), all using drive-by-wire. As a proof-of-concept, this build was planned to showcase the design’s modularity-- meaning that autonomy is a low priority for this particular vehicle iteration. For now we also nixed radio wireless because it might interfere with the braking mechanism. That left us with mechanical linkages to connect the module and platform, so the pedals in the module will have drive-pins to activate pistons in the platform. This linkage is also easily disengaged during a module swap (the platform will have handles to manually maneuver it under the modules, in addition to the module movement via the four actuators).
That’s one happy Marine! ^^
As you can see from the pictures, the welding of the construction module is almost complete, and the administrative module is at a similar state. Once the welding is done, then we’ll move on to the module build-up and reinstallation of the electrical system. All in all, we are very happy with the build progress, and seeing our fail-safe-and-fast decisions to borne out.
We want to hear your ideas on how to best balances the goals of demonstrating modularity and hackability while moving quickly with limited time and resources! Please share your thoughts below.