The development of a smaller, cheaper, and more robust class of mobile robots is the next step toward the integration of robots into human environments. To address this challenge, a scaled prototype of a ballbot (a mobile robot actively balanced on a sphere) was designed and built to investigate its various design requirements and performance limitations. The current design includes a body mounted on the rubber inner-core of a volleyball, which is actuated in two axes. Wheels powered by two

more »

... W DC motors drive the ball in an inverse trackball configuration. Since the weight of the robot sits atop a sphere, the system acts as a dual-axis inverted-pendulum. An inertial measurement unit (IMU) provides measurements of the tilt of the body to a microcontroller, which then implements a control algorithm to determine the motor speed necessary to dynamically stabilize the robot. A motor controller conveys this input to the motors via pulse-width modulation. Testing was developed to (a) qualitatively display stability, (b) determine the experimental proportional power relationship necessary to recover from varying tilt, and (c) discern the limiting factors in the design. The testing succeeded in providing a proof of concept as the robot did remain stable for 4±2 s. Furthermore, quantitative testing at varying initial tilts determined a linear relationship between power and tilt angle. Limiting components of the design included drifting readings from the IMU, a simple control algorithm that prevented long-term stability, losses and inconsistencies in the drive train, and a minimum voltage output from the motor controller preventing control at small angles. These lessons learned in phase one of design and construction have helped to set future goals for the project including the integration of a filter(s) or an attitude and heading reference system (AHRS) to reduce error in tilt angle readings, the substitution of a more sensitive motor controller, and the optimization of the drive system to minimize losses; all of which would facilitate the development of a more sophisticated control algorithm and pave the way for the integration of actively-balanced robots into human environments.