Pneumatic Autonomous Robot
Pneumatic Autonomous Robot
Embedded Files
I was responsible for the mechanical side of this class project (MAE106). We had nine weeks to build a robot from scratch under strict regulations.
The robot had to be dimension-restricted, fully autonomous, and powered solely by compressed air. It was required to navigate the course using dead reckoning and an Arduino-powered compass.
During the competition, seven or more robots had to travel through a narrow trench simultaneously. Because the robot's cylindrical shape was tightly constrained (Fig. 2), I defined three key design goals to maximize our chances of success with minimal design flexibility:
Low Center of Gravity – to reduce the risk of rolling over during collisions
Efficient Powertrain – to ensure the robot could reach the target before running out of air
Terrain Traversal Capability – to handle uneven surfaces, since the course was not perfectly flat
IMG_9652.movLow Center of Gravity
Low Center of Gravity
Since the tire height was fixed by regulation, I could only lower internal components to improve stability. I designed the chassis floor to sit as low as possible—just 0.536 inches from the ground—while maintaining sufficient ground clearance. I also positioned the heavier battery as close to the ground as possible to help lower the center of gravity.To prevent electromagnetic interference from the solenoid to the magnetometer, I mounted the solenoid on the second deck, away from the sensor.
Efficient Powertrain
Efficient Powertrain
We considered three propulsion options for the robot: a hopper mechanism, a rack and pinion system, and a piston-crankshaft mechanism. I ruled out the piston-crankshaft due to our limited timeframe, and dismissed the hopper (which uses a piston as a leg to push the robot forward) because it was too inefficient. The rack and pinion system was ultimately chosen for its simplicity in manufacturing and ease of packaging within our constrained size and schedule.
To maximize the power stroke length and simplify the design, I mounted the piston and rack-and-pinion assembly across the floor of the chassis. We also experimented with 3D-printed components to determine the optimal gear ratio, balancing torque and stroke efficiency, especially given the limited traction provided by the skateboard wheels.
Terrain Traversal Capability and 3D printing
Terrain Traversal Capability and 3D printing
With the robot sitting at such a low ride height, terrain traversal became a challenge. To address this, I designed the chassis with a negative rake—raising the front end to provide greater ground clearance for navigating dips, while placing more weight over the rear driving wheels to improve traction.
Aside from the wooden planks used for the floor, all components were 3D printed using PETG. This material choice allowed for rapid prototyping and structural flexibility. I also experimented with wall loops and sparse infill percentage in the front wheel hubs and axles—components that supported roughly 50% of the robot’s weight. This gave the front end just enough flex to absorb impact when crossing uneven terrain.
IMG_8637.MOV
Outcome
Outcome
We placed 2nd out of more than 40 entries in points despite a few human errors from our side. Our robot proved to be extremely robust, successfully meeting the initial design goals, as demonstrated in the video of one of the final heats of the competition. Thanks to the low center of gravity I prioritized in the design, our robot avoided rollover during collisions and remained upright even when opponents flipped over. It was also the lightest robot in the competition, weighing less than half of some of the heavier entries. Additionally, we were one of the fastest robots, due to our efficient rack-and-pinion powertrain and extensive testing to optimize the solenoid firing timing in relation to gear ratio (torque reduction).
Page updated
Report abuse