James Bruton, a UK-based engineer and YouTuber, has completed an omnidirectional bike using 3D printed components, aluminum extrusion, and a self-balancing control system. The design features two omni wheels mounted at 90 degrees to one another, enabling the vehicle to travel in any direction, including directly sideways. Bruton reused the front wheel from a one-wheel balancing robot built in April 2025 and the rear wheel from his earlier 2023 speeder bike project.
Aluminum 4040 T-slot extrusion forms the chassis, connected with sloped and right-angle brackets to hold the reused robot fork assembly in front and a belt-driven motor support in back. Axle holders and drivetrain pulleys were fabricated using LulzBot 3D printers, including the TAZ Workhorse and Mini 3. Bruton printed large structural components with 1.2 mm nozzles for faster build times. Polymaker’s PolyMax PLA was used for structural parts, while PolyLite Pro filament supported the electronics enclosure. Rear wheel hubs were made from plywood due to size limitations on the printer bed.
The drivetrain includes a two-stage belt reduction system. The first stage delivers a 3:1 torque increase using an intermediate pulley, followed by an HTD8-profile belt connecting to a large final pulley on the rear wheel. The rear drive system is powered by a single ODrive S1 servo motor kit. ODrive, a company specializing in open-source motor controllers, supplies the brushless motor and encoder assembly capable of delivering up to 2 kW. The system’s tension can be adjusted by sliding the motor mount along the extrusion frame, securing it with braced aluminum plates.
A PID controller maintains balance using data from a BNO086 inertial measurement unit. A Teensy 4 microcontroller processes roll data and regulates wheel torque accordingly. The control panel houses an emergency stop, start button, voltage monitors, trim adjustment, and a spirit level for orientation reference. A 500-amp contactor isolates motor power. Two pairs of 6S lithium polymer batteries in series supply 50V to the motor drivers and 12V to auxiliary systems.
Bruton replaced previous twist-grip input systems with two three-axis joysticks. The right-hand joystick handles forward, backward, and lateral movement. The left-hand stick controls rotation. Analog signals are processed through a smoothing filter adjustable via a control knob, allowing gradual deceleration when the joystick is released. Bruton observed that the filter improves ride quality by softening sudden changes in input.
Joystick input alters the PID controller’s setpoint, which defaults to 0°, representing upright balance. Adjustments to this setpoint allow the bike to lean and move simultaneously, similar to how a rider leans into turns. Bruton fine-tuned the PID parameters through trial and error, increasing the integral term to ensure that prolonged lean angles lead to increased motor response. He used minimal derivative correction to reduce overshooting and oscillations.
Initial testing showed that the rear omni wheel alone could stabilize the frame. When both wheels were active, the bike achieved omnidirectional movement. However, steering exposed torque imbalances. The rear wheel generated significantly more rotational force than the smaller front unit. When Bruton attempted to rotate in place, the chassis tipped under load, causing the front wheel to overcorrect and destabilize the frame. To compensate, he leaned opposite the intended turn direction, counteracting the chassis torque.
To simplify control, Bruton reconfigured the bike for reverse riding. He mounted the joystick interface at the rear and reversed stick mappings. With his weight now positioned over the larger rear wheel, torque disturbances diminished and steering response improved. “Now I can turn just fine… just lean the way I want to go and off we go,” he noted after reorienting the controls.
Despite mechanical success, the system has limitations. The small rollers on the front wheel require high current—around 20 amps at 50 volts even without load—which limits efficiency. The rear motor provides more torque than the front can compensate for, restricting balanced rotation at higher speeds. Bruton found that riding backward produced better control, as his center of mass aligned with the more powerful wheel.
All project files, including CAD designs and control firmware, are available on GitHub: https://github.com/XRobots/TwoOmniWheelBike.
Ready to discover who won the 2024 3D Printing Industry Awards?
Subscribe to the 3D Printing Industry newsletter to stay updated with the latest news and insights.
Featured photo shows CAD model of the omnidirectional bike. Image via James Bruton.
