People have dreamed of machines, which would free them from unpleasant, dull, dirty and dangerous tasks and work for them as servants, for centuries if not millennia. Service robots seem to finally let these dreams co...
ISBN:
(数字)9783642251160
ISBN:
(纸本)9783642251153
People have dreamed of machines, which would free them from unpleasant, dull, dirty and dangerous tasks and work for them as servants, for centuries if not millennia. Service robots seem to finally let these dreams come true. But where are all these robots that eventually serve us all day long, day for day? A few service robots have entered the market: domestic and professional cleaning robots, lawnmowers, milking robots, or entertainment robots. Some of these robots look more like toys or gadgets rather than real robots. But where is the rest? This is a question, which is asked not only by customers, but also by service providers, care organizations, politicians, and funding agencies. The answer is not very satisfying. Todays service robots have their problems operating in everyday environments. This is by far more challenging than operating an industrial robot behind a fence. There is a comprehensive list of technical and scientific problems, which still need to be solved. To advance the state of the art in service robotics towards robots, which are capable of operating in an everyday environment, was the major objective of the DESIRE project (Deutsche Service Robotik Initiative Germany Service robotics Initiative) funded by the German Ministry of Education and Research (BMBF) under grant no. 01IME01A. This book offers a sample of the results achieved in DESIRE.
We propose factorization as a concept to analyze and solve manipulation problems in unstructured environments. A factorization is a decomposition of the original problem into factors (sub-problems), each of which can ...
详细信息
ISBN:
(纸本)9783642194566
We propose factorization as a concept to analyze and solve manipulation problems in unstructured environments. A factorization is a decomposition of the original problem into factors (sub-problems), each of which can be solved much more easily than the original problem. The appropriate composition of these factors results in a robust and efficient solution to the original problem. Our assumption is that manipulation problems live in lower-dimensional subspaces of the high-dimensional state space associated with unstructured environments. A factorization identifies these subspaces and therefore permits finding simple and robust solutions to the factors. In this paper, we examine the effects of factorization in the context of our recent work on manipulating articulated objects in unstructured environments.
A primary challenge in the field of reconfigurable robotics is scaling down the size of individual robotic modules. We present a novel set of permanent magnet modules that are 900 mu m x 900 mu m x 270 mu m in size, c...
详细信息
ISBN:
(纸本)9783642194566
A primary challenge in the field of reconfigurable robotics is scaling down the size of individual robotic modules. We present a novel set of permanent magnet modules that are 900 mu m x 900 mu m x 270 mu m in size, called Mag-mu Mods, for use in a reconfigurable micro-system. The module is actuated by oscillating external magnetic fields less than 5 mT in strength, and is capable of locomoting on a 2-D surface. Multiple modules can be controlled by using an electrostatic anchoring surface, which can selectively prevent specific modules from being driven by the external field while allowing others to move freely. We address the challenges of both assembling and disassembling two modules. Assembly is performed by bringing two modules sufficiently close that their magnetic attraction causes them to combine. Disassembly is performed by electrostatically anchoring one module to the surface, and applying magnetic forces or torques from external sources to separate the unanchored module.
As robot bodies become more capable, the motivation grows to better coordinate them-whether multiple limbs attached to a body or multiple bodies assigned to a task. This paper introduces a new formalism for coordinati...
详细信息
ISBN:
(纸本)9783642194566
As robot bodies become more capable, the motivation grows to better coordinate them-whether multiple limbs attached to a body or multiple bodies assigned to a task. This paper introduces a new formalism for coordination of periodic tasks, with specific application to gait transitions for legged platforms. Specifically, we make modest use of classical group theory to replace combinatorial search and optimization with a computationally simpler and conceptually more straightforward appeal to elementary algebra. We decompose the space of all periodic legged gaits into a cellular complex indexed using "Young Tableaux", making transparent the proximity to steady state orbits and the neighborhood structure. We encounter the simple task of transitioning between these gaits while locomoting over level ground. Toward that end, we arrange a family of dynamical reference generators over the "Gait Complex" and construct automated coordination controllers to force the legged system to converge to a specified cell's gait, while assessing the relative static stability of gaits by approximating their stability margin via transit through a "Stance Complex". To integrate these two different constructs-the Gait Complex describing possible gaits, the Stance Complex defining safe locomotion-we utilize our compositional lexicon to plan switching policies for a hybrid control approach. Results include automated gait transitions for a variety of useful gaits, shown via tests on a hexapedal robot.
This paper addresses the challenge of developing robots that map and navigate autonomously in real world, dynamic environments throughout the robot's entire lifetime - the problem of lifelong navigation. Static ma...
详细信息
ISBN:
(纸本)9783642194566
This paper addresses the challenge of developing robots that map and navigate autonomously in real world, dynamic environments throughout the robot's entire lifetime - the problem of lifelong navigation. Static mapping algorithms can produce highly accurate maps, but have found few applications in real environments that are in constant flux. Environments change in many ways: both rapidly and gradually, transiently and permanently, geometrically and in appearance. This paper demonstrates a biologically inspired navigation algorithm, RatSLAM, that uses principles found in rodent neural circuits. The algorithm is demonstrated in an office delivery challenge where the robot was required to perform mock deliveries to goal locations in two different buildings. The robot successfully completed 1177 out of 1178 navigation trials over 37 hours of around the clock operation spread over 11 days.
Task demonstration is one effective technique for developing robot motion control policies. As tasks become more complex, however, demonstration can become more difficult. In this work we introduce a technique that us...
详细信息
ISBN:
(纸本)9783642194566
Task demonstration is one effective technique for developing robot motion control policies. As tasks become more complex, however, demonstration can become more difficult. In this work we introduce a technique that uses corrective human feedback to build a policy able to perform an undemonstrated task from simpler policies learned from demonstration. Our algorithm first evaluates and corrects the execution of motion primitive policies learned from demonstration. The algorithm next corrects and enables the execution of a larger task built from these primitives. Within a simulated robot motion control domain, we validate that a policy for an undemonstrated task is successfully built from motion primitives learned from demonstration under our approach. We show feedback to both aid and enable policy development, improving policy performance in success, speed and efficiency.
Achieving high performance autonomous navigation is a central goal of field robotics. Efficient navigation by a mobile robot depends not only on the individual performance of perception and planning systems, but on ho...
详细信息
ISBN:
(纸本)9783642194566
Achieving high performance autonomous navigation is a central goal of field robotics. Efficient navigation by a mobile robot depends not only on the individual performance of perception and planning systems, but on how well these systems are coupled. When the perception problem is clearly defined, as in well structured environments, this coupling (in the form of a cost function) is also well defined. However, as environments become less structured and more difficult to interpret, more complex cost functions are required, increasing the difficulty of their design. Recently, a class of machine learning techniques has been developed that rely upon expert demonstration to develop a function mapping perceptual data to costs. These algorithms choose the cost function such that the robot's planned behavior mimics an expert's demonstration as closely as possible. In this work, we extend these methods to address the challenge of dynamic and incomplete online perceptual data, as well as noisy and imperfect expert demonstration. We validate our approach on a large scale outdoor robot with hundreds of kilometers of autonomous navigation through complex natural terrains.
The topic of this paper is large-scale mapping and localization from images. We first describe recent progress in obtaining large-scale 3D visual maps from images only. Our approach consists of a multi-stage processin...
详细信息
ISBN:
(纸本)9783642194566
The topic of this paper is large-scale mapping and localization from images. We first describe recent progress in obtaining large-scale 3D visual maps from images only. Our approach consists of a multi-stage processing pipeline, which can process a recorded video stream in real-time on standard PC hardware by leveraging the computational power of the graphics processor. The output of this pipeline is a detailed textured 3D model of the recorded area. The approach is demonstrated on video data recorded in Chapel Hill containing more than a million frames. While for these results GPS and inertial sensor data was used, we further explore the possibility to extract the necessary information for consistent 3D mapping over larger areas from images only. In particular, we discuss our recent work focusing on estimating the absolute scale of motion from images as well as finding intersections where the camera path crosses itself to effectively close loops in the mapping process. For this purpose we introduce viewpoint-invariant patches (VIP) as a new 3D feature that we extract from 3D models locally computed from the video sequence. These 3D features have important advantages with respect to traditional 2D SIFT features such as much stronger viewpoint-invariance, a relative pose hypothesis from a single match and a hierarchical matching scheme robust to repetitive structures. In addition, we also briefly discuss some additional work related to absolute scale estimation and multi-camera calibration.
This paper addresses the cooperative localization and visual mapping problem for multiple aerial and ground robots. We propose the use of heterogeneous visual landmarks, points and line segments. A large-scale SLAM al...
详细信息
ISBN:
(纸本)9783642194566
This paper addresses the cooperative localization and visual mapping problem for multiple aerial and ground robots. We propose the use of heterogeneous visual landmarks, points and line segments. A large-scale SLAM algorithm is generalized to manage multiple robots, in which a global graph maintains the topological relationships between a series of local sub-maps built by the different robots. Only single camera setups are considered: in order to achieve undelayed initialization, we present a novel parametrization for lines based on anchored Plucker coordinates, to which we add extensible endpoints to enhance their representativeness. The built maps combine such lines with 3D points parametrized in inverse-depth. The overall approach is evaluated with real-data taken with a helicopter and a ground rover in an abandoned village.
暂无评论