UW engineers build palm-sized robot that hops, rights itself

A palm-sized machine capable of clearing standard stair steps, stabilizing after landing, and executing subsequent jumps has been engineered by University of Washington teams. The device, named DirectHop, investigates whether hopping mechanics provide small robots with a viable method for traversing obstacles without relying on flight capabilities.

The prototype relies on a compact electric motor paired with three folding legs to generate repeated leaps. Researchers control the jump height with single-centimeter precision by adjusting the electrical current delivered to the motor.

Direct motor propulsion mechanism

While many hopping robots utilize flea-inspired spring mechanisms, these systems often struggle with distance control and become difficult to manufacture at smaller scales. DirectHop bypasses springs entirely. Its motor connects via fishing line to a vertical tower structure on the robot’s body. Acceleration causes the motor to wind the line and climb the tower, lifting the robot into the air.

Three hinged legs extend during the ascent to maintain alignment between the motor and the robot’s foot. "We found that a very small electric motor can accelerate fast enough to power a jump directly, the way a frog’s leg muscles directly power its hop," stated Hanquan (John) Wang, lead author and UW graduate research assistant in mechanical engineering.

"We also discovered that by adjusting the current going to the motor, we could make the robot jump specific distances," Wang added.

Self-righting and recovery systems

Landing stability poses a significant challenge for robots requiring multiple hops. Because DirectHop tends to tumble mid-air, the team integrated a roll cage designed to support the unit when it lands on its side. This feature draws inspiration from box turtle shells, which assist animals in rolling upright.

Researchers achieved a 90% success rate in righting the robot by moving the motor back down the tower. This action shifts the center of gravity, causing the robot to roll onto its foot and prepare for the next jump.

"In my mind, the three hardest challenges for hopping robots are the ability to vary jump length, self-right and reload for multiple hops," said Sawyer Fuller, senior author and UW associate professor of mechanical engineering.

Future autonomy and applications

Current iterations of DirectHop require wired power and lack steering capabilities. However, Wang and Fuller are developing onboard solar cells, batteries, and vibration motors for turning. They are also integrating retractable feet to adjust hop angles and cameras with electronics for navigation.

The ultimate goal involves autonomous stair climbing, where the robot identifies a step, calculates the necessary jump, positions itself, and repeats the sequence after landing. Wang estimates that a fully equipped version could cost approximately $10 to produce.

If technical hurdles are resolved, swarms of these inexpensive robots could potentially inspect oil refineries, monitor agricultural sites, or explore planetary surfaces. Such deployments would allow them to operate alongside larger machines in environments where losing an individual unit does not compromise the overall mission.