Robots are an integral part of contemporary society with ever increasing use in resource collection, industrial, military, and exploration-based applications. The degree of human intervention necessary for a robot to successfully operate varies from system to system and depends heavily on the application, resources available, and knowledge of the operational environment. A system that can operate either autonomously or under control of a human operator provides the flexibility to adapt to the requirements of a specific mission and to unexpected situations as they are encountered. Similarly, a platform that can be quickly and easily modified is capable of being utilized in a larger number of applications. This thesis describes a generalized computational architecture for a robotic platform that allows the system to be either teleoperated by a remote human operator or completely autonomous. In addition, the architecture is designed specifically to support a modular platform upon which various modules can be attached or removed to provide the functionality needed to complete specific tasks. This architecture was implemented on the University of Alabama Modular Autonomous Robotic Terrestrial Explorer (MARTE) platform as part of the 2014 NASA Robotic Mining Competition to ultimately result in a robot capable of collecting and delivering regolith in a simulated lunar or Martian environment. MARTE was successfully used at the competition and demonstrated the capabilities of the architecture both during testing and in the competition environment.