Contingency basing presents a planner with numerous design decisions driven by multiple design criteria such as the number of soldiers, base permanency, base location, and other factors. The operational environment of the base is not static either; design requirements change as the mission changes. In this work, we introduce a model-based systems engineering approach to elicit design and operational needs while dealing with the design complexity of constructing a contingency base. The model includes the key facility types that can make a contingency base, interactions between facility types, and required utilities for each facility type. The model elements are kept at an abstract level so the details can be altered as required by the customer needs. Pairing the model with an external analysis tool allows for quick development and testing. Properties of the facility types can be altered either in the model or the analysis tool, and reflected in both. Using the model-based systems engineering concepts of reusability, these elements can be saved and re-used in future base designs allowing for a rapid and adaptable design process. In addition, the sharing of information visually with Object Management Group's Systems Modeling Language TM diagrams enhances the ability to collaborate with nonengineering subject matter experts within the design domain. By graphically showing the conditions and layout of the proposed contingency base, Department of Defense personnel not trained in modeling and simulation were able to interact with the engineering designs and identify gaps in the proposed architecture. C⃝ 2015 Wiley Periodicals, Inc. Syst Eng 18: 162-177, 2015

No abstract

No abstract

Designing controllers for skid-steered wheeled robots is complex due to the interaction of the tires with the ground and wheel slip due to the skid-steer driving mechanism, leading to nonlinear dynamics. Due to the recent success of reinforcement learning algorithms for mobile robot control, the Deep Deterministic Policy Gradients (DDPG) was successfully implemented and an algorithm was designed for continuous control problems. The complex dynamics of the vehicle model were dealt with and the advantages of deep neural networks were leveraged for their generalizability. Reinforcement learning was used to gather information and train the agent in an unsupervised manner. The performance of the trained policy on the six degrees of freedom dynamic model simulation was demonstrated with ground force interactions. The system met the requirement to stay within the distance of half the vehicle width from reference paths.