A small-scale agile wall climbing robot able to navigate on smooth surfaces of any orientation, including vertical and inverted surfaces, which uses adhesive elastomer materials for attachment. The robot uses two actuated legs with rotary motion and two passive revolute joints at each foot the robot can climb and steer in any orientation.

Carnegie Mellon


Mass100 g
Length13 cm
Height5 cm
Width12.3 cm
Turning Radius (Rt)10 cm
Turning Circle Radius (Rtc)11.15 cm
Step Length (dstep)2.6 cm
Adhesive Area (per foot)1.1 cm2
Speed6 cm/s
Turning Speed37.5 0/s
Powerlithium ion polymer battery
Motors2 Sanyo 12GA-N4s


Waalbot is intended for inspection and surveillance applications and, ultimately, space missions.

The robot uses two actuated legs with rotary motion and two passive revolute joints at each foot the robot can climb and steer in any orientation. The Waalbot has onboard power, sensing, computing, and wireless communication which allow for semi-autonomous operation.

The robot can climb 90 degree slopes at a speed of 6 cm/s and steer to any angle. The robot uses polymer adhesive material (Smooth-On Vytaflex 10) as an adhesive.

A gear motor’s output shaft is connected to a triangular shaped leg, where each point of the triangle holds a foot assembly on a revolute ankle joint. The ankle joint is spring loaded to always return to the forward position. On the end of the foot is a passive revolute joint which connects a foot pad that  holds the adhesive material. This distal joint allows the robot to steer.

When moving forward, the two legs are synchronized and step in unison. As the motors turn, the tail of the robot presses against the surface and the legs rotate forward. The two feet which are
sticking to the surface support the weight of the robot.The robot now has 5 contact points with the surface; 2 feet on each side and the tail. The motor torque provides an internal moment which presses the front feet onto the surface while pulling the rear feet away from the surface.

The robot is controlled by a PIC microcontroller (PIC16F737) and is able to perform pre-programmed actions such as climbing and turning. The speed is controlled with feedback from foot position sensors. Limit switches are triggered when the legs are aligned such that only one foot on each side is
contacting the surface. This keeps the robot’s speed synchronized by stopping one of the motors until
the other motor catches up.

Infrared (IR) RC5 communication is used to command the robot to climb straight, stop, or turn.


Describes the project and its design. Goes into the attachment mechanism, and adhesive requirements. Describes the fabrication, steering, experiments, and finally the conclusion.

Michael P. Murphy, William Tso, et al. - EEE/RSJ International Conference on Intelligent Robots and Systems, 2006.

Describes the project, and both designs. Discusses the dry adhesion, the prototype, experiments, and the conclusion.

Ozgur Unver, Michael P. Murphy, et al. - September 2005.

Wevolver 2023