This paper explores a two state rover concept called the Transforming Roving-Rolling Explorer (TRREx). The first state allows the rover to travel like a conventional 6-wheeled rover. The second state is a sphere to permit faster descent of steep inclines. Performance of this concept is compared to a traditional rocker-bogie (RB) architecture using hi-fidelity simulations in Webots. Results show that for missions involving very rugged terrain, or a considerable amount of downhill travel, the TRREx outperforms the rocker-bogie. Locomotion of the TRREx system using a continuous shifting of the center of mass through “actuated rolling” is also explored. A dynamics model for a cylindrical representation of the rover is simulated to identify feasible configurations capable of generating and maintaining continuous rolling motion even on sandy terrain. Results show that in sufficiently benign terrain gradual inclines can be traversed with actuated rolling. This model allows for increased exploration of the problem's design space and assists in establishing parameters for an Earth prototype.

This paper presents a strategy for assessing reconfigurability at the architecture level in the design of a complex system to fulfill a chaotic objective. A hi-fidelity simulation environment is used to quickly run a myriad of test scenarios on a reconfigurable system and a more traditional alternative. Specifically, the design of a mobility system for a Mars rover is used in this assessment. Rovers provide an excellent illustration of the issue; 1) the mobility problem is one with many random variables, 2) rover architectures are complex enough that a make-it, break-it, fix-it approach is prohibitively expensive and time consuming, and 3) several rover architectures employing reconfigurability have been proposed or are already under development. The ultimate objective of this research is to compare complex systems across a widely diverse range of architectures and reconfigurations. This paper is a step toward that goal. Four roverstwo architectures using two size scales-are tested in twenty terrain challenges, and their performance is explored across various levels of ground traction, slope, and rock field density. By considering potential missions as a combination of the objectives of the terrain challenges, architecture selection is explored. Performance is the primary measure considered in this paper with cost, risk, and other considerations to be addressed in future work. I.

Reconfigurable system design is known to be a strategy that offers vast performance improvement over traditional designs. Extending from previous work, this paper presents a strategy for assessing the domination of a reconfigurable concept. By considering the performance demands placed upon a system at different segments of a mission, reconfigurable systems are shown in a multi-objective space as a collection of points. This is different than static designs which are typically represented as a single point. The principles of Pareto dominance are extended to describe the comparisons between systems in this space. A surrogate point is introduced to reduce calculation burden and provide a foundation for the development of a necessary condition for dominance. This approach is then demonstrated on a case study of Mars exploration rovers where a traditional rocker-bogie architecture is compared to the two-state Transforming Rolling Roving Explorer architecture. These principles lay the groundwork for choosing between reconfigurable and static systems when multiple objectives are considered. I. IntroductionR econfigurability and flexibility are concepts associated with complex system design that offer large improvements in system performance and robustness. Reconfigurability is the ability of a system to change "configurations … repeatedly and reversibly," 1 and benefits include multi-ability (operating at multiple performance points in the design space non-simultaneously), evolvability (changes to meet new demands), and survivability (changes to maintain functionality despite component failures). The goal of this paper is to explore the advantages of multi-ability when initially selecting system architecture. Specifically, this paper explores the architecture selection for a Mars rover.Tradespace visualization is a powerful tool for navigating the trade-offs inherent in system design, especially when multiple objectives are considered. This paper explores the challenges with, and establishes initial groundwork for, visualizing the complexity associated with reconfigurable systems in a multidimensional environment. Key challenges of this task relate to understanding how performance domination in a multi-objective space can be extended to a reconfigurable system, and how this domination may be explored visually. Solving these challenges will allow for the development of fundamental rules for evaluating reconfigurable architectures that can then be used to facilitate the application of multi-objective optimization techniques toward reconfigurable designs.Research over the last decade has seen an increased interest in the optimization of reconfigurable systems. Unmanned aerial vehicles have been a popular topic in particular. Work in this area has conducted sensitivity studies 2 , explored concept embodiment 3 , and developed analyses engines capable of spanning multiple disciplines 4-6 . Reconfigurable UAVs have also been envisioned with offline reconfigurations where the UAV is changed between missions by swapping out wi...

A seven-step framework for sorting proposed concepts of system changes / reconfigurations is presented that seeks to characterize the overall ramifications on system architecture. This framework is intended for use immediately following a concept generation phase. The framework uses three simple questions: “What level of the system design does this concept apply to?” “What levels of the system design does the concept impact?” and “What is the severity of this impact?” A flowchart leads the designer through these questions and assigns each concept a classification from one to five based on the answers. Class one concepts have little to no impact on the rest of the system architecture. They can be included with little fear of massive change propagation and system redesign. Class five concepts carry large changes to system architecture and therefore should be included only if they can be shown to be highly beneficial, or if there remains enough design freedom such that the cost of changing the system architecture is minimal. Meanwhile, class five concepts are likely to have much higher potential to create revolutionary design. A case study is used to demonstrate the application of the sorting framework in the context of a Mars rover mission. Several example concepts are provided to illustrate key insights from the case study. Convergence of the framework is explored by comparing the authors’ results to a second test done by a new design team.

The ultimate goal of this research is to provide computer based educational software that exposes engineering students to design tradeoffs early in their undergraduate experience. This paper investigates two feedback elements for their ability to enhance those students’ understanding of the tradeoffs inherent in a water rocket propulsion design problem: 1) a Latin hypercube sample that allows the student to select a starting point and 2) sensitivity values that displayed local gradient information. Assessments are made using data logged during the students’ interaction with the software and a series of quizzes performed throughout the study. The results indicate that the sensitivity information improves the students’ ability to locate designs with good performance, while the Latin hypercube adversely affects the students’ ability to visualize the objective space.