| Robot configuration design is hampered by the lack of established, well-known design rules, and designers cannot easily grasp the design space and the impact of design variables on robot performance. A human can only design and evaluate several candidate configurations, though there may be thousands of competitive designs that should be investigated. In contrast, an automated synthesis method can create and evaluate tens of thousands of designs without relying on previous experience or design rules.;This thesis develops Darwin2K, an extensible, automated system for robot configuration synthesis. The focus is on core capabilities for robot synthesis: a flexible, extensible, and effective synthesis algorithm, simulation capabilities, and representation of robots and their properties. The system can synthesize and optimize kinematics, dynamics, structural geometry, actuator selection, and task and control parameters for a wide range of robots.;Darwin2K uses an evolutionary algorithm with two new multiobjective selection procedures that are applicable to other evolutionary domains. The EA can effectively optimize multiple performance objectives while satisfying multiple constraints, and can generate a range of trade-off solutions. A novel robot representation, the parameterized module configuration graph, enables efficient and extensible synthesis of mobile robots, of single, multiple and bifurcating manipulators, and of modular and monolithic robots.;Task-specific simulation provides the synthesizer with performance measurements for each robot. Darwin2K can automatically derive dynamic equations for each robot, enabling dynamic simulation to be used during synthesis for the first time. Darwin2K's software toolkit includes a library of robot subassemblies and components, numerous performance metrics, and simulation capabilities such as collision detection and estimation of link deflection. An extensible, object-oriented software architecture allows new toolkit components to be added without impacting the synthesizer.;Darwin2K's synthesis algorithm, simulation capabilities, and extensible architecture combine to allow synthesis of robots for a wide range of tasks. Results are presented for nearly 150 synthesis experiments for six different applications, including synthesis of a free-flying robot with multiple manipulators and a robot for zero-gravity truss walking. The synthesis system and results represent a significant advance in the state-of-the-art in automated synthesis for robotics. |
