One of the problems affecting robot operators' spatial awareness involves their ability to infer a robot's location based on the views from on-board cameras and other electro-optic systems. To understand the vehicle's location, operators typically need to translate images from a vehicle's camera into some other coordinates, such as a location on a map. This translation requires operators to relate the view by mentally rotating it along a number of axes, a task that is both attention-demanding and workload-intensive, and one that is likely affected by individual differences in operator spatial abilities. Because building and maintaining spatial awareness is attention-demanding and workload-intensive, any variable that changes operator workload and attention should be investigated for its effects on operator spatial awareness. One of these variables is the use of automation (i.e., assigning functions to the robot). According to Malleable Attentional Resource Theory (MART), variation in workload across levels of automation affects an operator's attentional capacity to process critical cues like those that enable an operator to understand the robot's past, current, and future location. The study reported here focused on performance aspects of human-robot interaction involving ground robots (i.e., unmanned ground vehicles, or UGVs) during reconnaissance tasks. In particular, this study examined how differences in operator spatial ability and in operator workload and attention interacted to affect spatial awareness during human-robot interaction (HRI). Operator spatial abilities were systematically manipulated through the use of mental transformation training. Additionally, operator workload and attention were manipulated via the use of three different levels of automation (i.e., manual control, decision support, and full automation). Operator spatial awareness was measured by the size of errors made by the operators, when they were tasked to infer the robot's location from on-board camera views at three different points in a sequence of robot movements through a simulated military operation in urban terrain (MOUT) environment. The results showed that mental transformation training increased two areas of spatial ability, namely mental rotation and spatial visualization. Further, spatial ability in these two areas predicted performance in vehicle localization during the reconnaissance task. Finally, assistive automation showed a benefit with respect to operator workload, situation awareness, and subsequently performance. Together, the results of the study have implications with respect to the design of robots, function allocation between robots and operators, and training for spatial ability. Future research should investigate the interactive effects on operator spatial awareness of spatial ability, spatial ability training, and other variables affecting operator workload and attention.

The robotics community has succeeded in creating remarkable machines and task-level programming tools, but arguably has failed to apply sophisticated autonomous machines to sophisticated tasks. One reason is that this combination leads to prohibitive complexity. Biological systems provide many examples of integrated systems that combine high-performance and flexibility, with logically-organized low-level control. Sophisticated organisms have evolved that depend on physical dexterity to thrive in a particular ecological niche while mitigating computational and behavioral complexity. This dissertation investigates the potential for a new kind of hybrid robotic design process. A design for performance that combines mechanical dexterity with low-level embedded firmware that organizes behavior and facilitates programming at a higher level. I propose that dexterous machines can incorporate embedded firmware that express the "aptitudes'' implicit in the design of the robot and hierarchically organize the behavior of the system for programming. This is a win-win situation where the quality of the embedded firmware determines how efficiently programmers (autonomous learning algorithms or human programmers) can construct control programs that are robust, flexible, and respond gracefully to unanticipated circumstances. This dissertation introduces the uBot-5---a mobile manipulator concept for human environments that provides dexterous modes for mobility and manipulation and control firmware that organizes these behavioral modes locally for use by applications code. Postural control underlies the uniform treatment of several mobility modes that engage different combinations of sensory and motor resources. The result is a platform for studying "whole-body'' control strategies that can be applied jointly to simultaneous mobility and manipulation objectives. The thesis examines the specification and development of both: (1) a dexterous robot for unstructured environments, and (2) the embedded firmware that organizes dexterous behavior for mobility and manipulation tasks. Integrated solutions are proposed that control transitions between postural "modes'' and provide a logically organized dexterous behavior hierarchy. Firmware programming can also be used to construct an efficient API for user programming and autonomous machine learning. My goal is to contribute technologies that can support new robotic applications in our culture that require fully integrated dexterous robots in unstructured environments. Personal robotics is an important emerging application that depends on seamlessly integrated and sophisticated machines, controllers, and adaptability. Logically organized representations for use in task-level application development are critical to pull this off. The impact of such technology could be significant---with applications that include healthcare and telemedicine, exploration, emergency response, logistics, and flexible manufacturing.

Legged robots have the potential to cover terrain not accessible to wheel-based robots and vehicles. This makes them better suited to perform tasks, such as search and rescue, in real-world unstructured environments. Pneumatically-actuated, compliant robots are also more suited than their rigid counterparts to work in real-world unstructured environments with humans where unintentional contact may occur. This thesis seeks to combine the benefits of these two type of robots by implementing design methods to aid in the design choice of a 16 degree of freedom (DoF) compliant, continuum-joint quadruped. This work focuses on the design optimization, especially the definition of design metrics, for this type of robot. The work also includes the construction and closed-loop control of a four-DoF continuum-joint leg used to validate design methods.We define design metrics for legged robot metrics that evaluate their ability to traverse unstructured terrain, carry payloads, find stable footholds, and move in desired directions. These design metrics require a sampling of a legged-robot's complete configuration space. For high-DoF robots, such as the 16-DoF in evaluated in this work, the evaluation of these metrics become intractable with contemporary computing power. Therefore, we present methods that can be used to simplify and approximate these metrics. These approximations have been validated on a simulated four-DoF legged robot where they can tractably be compared against their full counterparts.Using the approximations of the defined metrics, we have performed a multi-objective design optimization to investigate the ten-dimensional design space of a 16-DoF compliant, continuum-joint quadruped. The design variables used include leg link geometry, robot base dimensions, and the leg mount angles. We have used an evolutionary algorithm as our optimization method which converged on a Pareto front of optimal designs. From these set of designs, we are able to identify the trade-offs and design differences between robots that perform well in each of the different design metrics. Because of our approximation of the metrics, we were able to perform this optimization on a supercomputer with 28 cores in less than 40 hours.We have constructed a 1.3 m long continuum-joint leg from one of the resulting quadruped designs of the optimization. We have implemented configuration estimation and control and force control on this leg to evaluate the leg payload capability. Using these controllers, we have conducted an experiment to compare the leg's ability to provide downward force in comparison with its theoretical payload capabilities. We then demonstrated how the torque model used in the calculation of payload capabilities can accurately calculate trends in force output from the leg.

Multi-robot systems of autonomous mobile robots offer many benefits but also many challenges. This work addresses collision avoidance of robots solving continuous problems in known environments. The approach to handling collision avoidance is here to enhance a motion planning method for single-robot systems to account for auxiliary robots. A few assumptions are made to put the focus of the work on path planning, rather than on localization. A method, based on exact cell decomposition and extended with a few rules, was developed and its consistency was proven. The method is divided into two steps: path planning, which is off-line, and path monitoring, which is on-line. This work also introduces the notion ofpath obstacle, an essential tool for this kind of path planning with many robots. Furthermore, an implementation was performed on a system of omni-directional robots and tested in simulations and experiments. The implementation practices centralized control, by letting an additional computer handle the motion planning, to relieve the robots of strenuous computations. A few drawbacks with the method are stressed, and the characteristics of problems that the method is suitable for are presented. / QC 20100705

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.