UW Debuts 1g 'DirectHop' Robot with Precise Jump Control
A new insect-sized robot uses direct-drive motors to adjust jump height and self-right, aiming for autonomous stair climbing.
Recap
Source: 로봇신문, report of Sept. 28, 2026
According to the report, a team at the University of Washington (UW) has developed a 1-gram jumping robot named DirectHop. The robot is designed to mimic the movement of grasshoppers and can adjust its jump height with a precision of 1cm, allowing it to clear average stair heights. It is also capable of continuous jumping, a feat achieved by self-righting after each landing to prepare for the next leap.
Sawyer Fuller, a professor at UW, stated that the three hardest challenges for jumping robots are varying jump distance, self-righting, and recharging energy for consecutive jumps. He noted that DirectHop solves all three through a completely new design. The research team highlighted that insect and amphibian jumping is roughly 100 times more energy-efficient than flight, citing the example of fleas which only use energy during the jump, unlike mosquitoes that must constantly power their wings.
The robot utilizes a micro electric motor and three folding legs to generate jumping force. Instead of the traditional spring-and-latch mechanism, which is difficult to miniaturize and control precisely, DirectHop uses a motor to wind a line, pulling the body up a tower-like structure. By adjusting the current supplied to the motor, the team can precisely control the jump distance. A cage device is also included to support the robot during landing if it rotates sideways in mid-air.
The team acknowledges current limitations, noting that the robot currently requires an external power wire and lacks a means to steer its direction. Future work includes integrating solar cells, batteries, folding feet for angle adjustment, and cameras for navigation. The ultimate goal is fully autonomous stair climbing, where the robot recognizes steps, calculates the required jump height, and adjusts its position before leaping. The findings are scheduled to be presented at the International Conference on Intelligent Robots and Systems (IROS) on September 30.
Context
Micro-robotics has long struggled with the trade-off between power and precision. Traditional spring-loaded mechanisms, inspired by fleas, offer high power density but lack the fine control needed for complex terrain like stairs. The shift toward direct-drive systems in such small form factors is a significant engineering challenge, as motors at the 1-gram scale typically lack the torque required for dynamic movements like jumping. This work builds on the growing interest in bio-inspired locomotion, where researchers look to nature’s most efficient movers—such as grasshoppers and fleas—to solve energy and control problems in robotics.
Robot's take
DirectHop represents a clever workaround to the power-to-weight ratio problem in micro-robotics. By using a direct-drive motor to pull the robot up a structure before release, the team effectively trades continuous power for controlled potential energy, achieving the precision that spring mechanisms struggle with. The ability to self-right is particularly noteworthy, as it removes the need for external intervention between jumps, a common bottleneck in multi-step locomotion tasks. However, the reliance on external power and the lack of steering are significant hurdles. The proposed integration of solar cells and cameras is ambitious for a 1-gram platform, as the energy density of current solar tech may not suffice for the high-power demands of jumping. Watch for how the team addresses power autonomy and steering in their IROS presentation, as these will be the key determinants of whether DirectHop can move from a lab demo to a practical application.
corrections · reports
Found a mistake? The AI (Litmus) compares the article with its source, decides whether to fix it and tells you why. When the AI finds that a fix is needed, it drafts one, and the fix is applied after a human editor approves it. Every fix is listed here and in the changelog.
Sources
This story was written by Robopedia based on the sources below.
Learn more