One of the emerging challenges in mobile robotic is the development of devices capable of mobility using multiple modes of locomotion such as running over and under ceilings, climbing walls, and flying through building interiors. These robots will exhibit unprecedented mobility, but their development will require advances in the application of bio-inspired dynamic locomotion principles and the integration of smart material and manufacturing developments. Legged robots, such as the Sprawl and RHex families of robots have demonstrated that fast and robust motion over rough terrain can be achieved with simple control if the passive compliance and geometry of the legs is properly designed. Similar principles have led to the development of climbing robots that can scale natural environments such as trees and buildings, (RiSE), and rapid dynamic climbing at animal-like speeds (DynoClimber). Flying robots based on insects, mammals, and birds are also being developed for flight in tight, confined spaces. While robots have recently been developed that can move dynamically in each of these domains, no robot exists today that can operate in all three. In this project we aim to leverage recent developments in attachments mechanisms from the RiSE project3, and rapid vertical mobility4 and smart-material based variable structures developed in my lab to create small (50g) robots capable of motion in complex, indoor environments. One of the greatest challenges in creating this type of robotic system is the efficient exploitation of on-board power. For example, in running systems the efficiency of motion is greatly improved via cyclic storage and return of energy. Proper tuning of these spring elements is essential to robust performance. Evidence is beginning to mount that the adaptation of the leg spring stiffness in running can improve both the efficiency and the stability of the resulting motion. For example, recently lightweight variable complaint legs have been developed for a hexapedal robot to improve its performance as it runs over different terrains and under different loading conditions. This work, while successful, relied on small DC motors and complex mechanical systems to alter the leg geometry, adding weight and complexity and reducing the robustness of the design. In general, generating sufficient torque and transmitting power from commercial motor technology is difficult to implement on small, centimeter scale, robots. While running, climbing and flying can all utilize cyclic motions, the appropriate limb motions and passive properties for each mode of locomotion are distinct. Thus the passive properties of the limb suspension system and the geometry will need to be altered for each mode. At the scales that a combined device is desired (less than 100g), using multiple DC motors become infeasible. An early, 350g, prototype hybrid robot we have developed for rapid climbing on and gliding from walls. It can climb prepared surfaces at speeds over 10 cm/s and launch itself into a glide, but its inability to maneuver, change gaits, or switch from vertical to horizontal motion is largely due to limitations in using fixed airfoil wings and off-the-shelf actuator/transmission technologies. To address this problem we will integrate smart materials, such as shape memory alloy and polymer composites and electro-active polymers to the limbs of dynamic robots to provide the needed adaptability at the required scales. The fabrication techniques for multi-material complaint structures used by the Shape Deposition Manufacturing process will provide a framework for the integration of these active materials into small robots with flexible, collapsible wings.