Mobile Bimanual Manipulation Robot
Developed an untethered, low-cost bimanual mobile manipulator featuring a custom Tri-Bus power topology and structurally optimized chassis to support GPU-accelerated edge compute.
Tip: Use J / L to skip or F for fullscreen
Cutting the Cord
While open-source robotics has lowered the barrier to entry, achieving true untethered deployment introduces severe integration challenges. Compliant 3D-printed links waste motor torque, and simultaneous actuation of a 17-DoF system causes voltage collapse under high transient loads, leading to unpredictable compute brownouts during bimanual tasks.
Tri-Bus Power & Structural Optimization
To support an onboard NVIDIA Jetson Orin Nano, we engineered a custom Tri-Bus power topology with firmware-level current envelopes. This physically and logically isolates the compute module from the drivetrain's dynamic electrical loads.
Combined with a functionally graded "High-Shell" 3D-printing strategy that increases end-effector payload capacity by 67%, the system achieves highly reliable, zero-latency autonomy for complex tasks like VR teleoperation and learning-based policy execution.