MARTIN LI/ MEGet in touch
BACK TO PROJECTS

02 / ROBOTICS

Code meets hardware.

UC Irvine · MAE 106

ACADEMIC PROJECT · UNPAID

01 / PNEUMATIC MOBILE ROBOT

Controls, fabrication & integration.

Developed the main control program and fabricated chassis components for a three-person pneumatic robot project. Combined Arduino/C++ programming, calibration, and hands-on mechanical repairs to help turn a team design into working hardware.

My contribution

  • Control programming. Wrote the main Arduino/C++ program for the robot, with timed solenoid actuation, heading-based steering, and distance-based navigation.
  • Calibration and debugging. Adjusted control parameters with teammate support to bring the program and physical robot into alignment.
  • Chassis fabrication. Cut and sanded the upper and lower circular wooden plates and structural support columns.
  • Mechanical troubleshooting. Replaced failed supports made from 3D-printed parts and aluminum rods with wooden columns, helping restore the assembly.
Arduino / C++Feedback controlSensor integrationCalibrationWood fabricationMechanical troubleshooting

How the robot works

A compressed-air tire supplies a pneumatic cylinder. A rack-and-pinion mechanism drives two gear trains, using both the powered stroke and the spring return. The team used a 3:1 ratio on the powered side and 1:4 on the return side to balance speed and torque.

The control system combines a steering servo, magnetometer/IMU heading feedback, wheel-switch distance measurement, and a timed solenoid. The final program sequences straight travel and two 90-degree turns.

Team: Martin Li, Adam Diersing, and Logan Bigler. Logan led CAD and drivetrain design; Adam handled wiring, electrical work, test data, and final code fixes.

Read the team’s final report

TEAM TEST RESULT

6.77 seconds over 10 feet.

The best tested setting averaged 6.774 s across 10 runs, with a standard deviation of 0.156 s. Tests used a 40% solenoid duty cycle, a 750 ms period, 40 psi starting pressure, and a heading-control gain of 0.6.

This setting also showed the lowest variation among the tested configurations. It was selected for the final program.

Experimental testing and analysis: Adam Diersing. Results describe the team’s robot.

Report graph of mean time to travel ten feet versus solenoid duty cycle at a 750 millisecond period, with one-standard-deviation error bars
Team report: 750 ms period, ten trials per setting. Lower time is faster.

FROM THE PROJECT REPORT

Physical hardware and team design documentation from our MAE 106 final report.

Physical wheeled pneumatic robot with an upright tire, wooden chassis, and drive mechanism
01 / THE PHYSICAL ROBOTOur pneumatic robot, with the tire storing the compressed air that drives it.
Side view of the team’s CAD robot assembly and pneumatic drive
02 / CAD / SYSTEM LAYOUTThe team’s CAD assembly shows how the tire, cylinder, chassis, and gear trains fit together.
Angled CAD view showing rack-and-pinion gearing beneath the robot’s tire
03 / CAD / DRIVE MECHANISMA closer look at the rack-and-pinion drive and the gears underneath.
Team wiring diagram of Arduino Uno, magnetometer, servo, MOSFET, solenoid and power supplies
04 / WIRING DIAGRAMThe electrical layout connecting the Arduino, heading sensor, steering servo, and solenoid driver.

02 / EARLIER ROBOTICS WORK

Two-legged running robot.

In a separate UC Irvine project, I led an eight-person team through mechanical design, fabrication, and assembly of a two-legged running robot.

Modeled components and the assembly in SolidWorks, integrated 3D-printed and woodcut parts, and used hand tools, precision fasteners, and fit checks to refine the physical build.

Rocket projectCisco Antigua