The SEAS Robotics Program

Professor Pinhas Ben-Tzvi conducts research on the advanced mechanics and control of robotics and mechatronics systems, the design of intelligent autonomous systems, and the development of novel sensors and actuators for microrobotic systems, miniature mechatronic systems, and biomedical applications. In one of his robotics research activities, Professor Ben-Tzvi developed a new paradigm that generated a revolutionary improvement in mobile robot design: a hybrid system that combined locomotion (the ability to move) and manipulation (the ability to grasp, carry, etc.) and increased both the strength and resiliency of the mobile robot.

"The previous approach was to create two platforms for the robot, a mobility platform with a manipulation platform on top of it," explains Professor Ben-Tzvi. "The problem with these designs was that the degrees of freedom for mobility are used only for mobility and those for manipulation are used only for manipulation. These robots work outside, for example in search and rescue missions or earthquake sites, and it's highly likely that they will fall and flip over at some point and then break, so they're very limited. The hybrid mechanism mobile robot typically has half of the number of degrees of freedom, because its links are used simultaneously for both locomotion and manipulation; they're sharing the functionalities symbiotically and they're not as susceptible to breaking."

Under a grant from the Defense Advanced Research Projects Agency, Professor Ben-Tzvi's lab is now working on a new mobile robot design with the capability of operating autonomously rather than through wireless remote control. This is extremely difficult to do, because the infrastructure necessary for autonomy requires much greater computing capacity and speed, new sensing capabilities, and complex decision making algorithms in order to perceive the environment and act upon it.

Professor Ben-Tzvi has also proposed a new project, Self-Configuring Robotic Pieces (SCoRP), which he explains with a scenario. "We would design smaller mobile robots that can move around very effectively, like agents," he states. "They would be scattered all around, say in a collapsed building after an earthquake, and they would act as sensor agents collecting information through cameras and sensors, which they would share through wireless communication. When a need arises to actually act on the environment-for example, one of the robots finds a person in the rubble-then it calls the other robots, which swarm in from various locations and they self-assemble into the hybrid configuration needed for a particular task. We are going to pioneer this at GW," he says.