On Intelligent Robot Motion Planning via Learning

Automatic motion planning is one of the basic modules that are needed to increase robot intelligence and usability. Unfortunately, the inherent complexity of motion planning has rendered traditional search algorithms incapable of solving every problem in real time. To circumvent this difficulty, we explore the alternative of allowing human operators to participate in the problem solving process. By having the human operator teach during difficult motion planning episodes, the robot should be able to learn and improve its own motion planning capability and gradually reduce its reliance on the human operator. In this paper, we present such a learning framework in which both human and robot can cooperate to achieve real-time automatic motion planning. To enable a deeper understanding of the framework in terms of performance, we present it as a simple learning algorithm and provide theoretical analysis of its behavior. In particular, we characterize the situations in which learning is useful, and provide quantitative bounds to predict the necessary training time and the maximum achievable speedup in planning time.     

Learning Degradation-Invariant Representation for Robust Real-World Person Re-Identification

Person re-identification (Re-ID) in real-world scenarios suffers from various degradations, e.g., low resolution, weak lighting, and bad weather. These degradations hinders identity feature learning and significantly degrades Re-ID performance. To address these issues, in this paper, we propose a degradation invariance learning framework for robust person Re-ID. Concretely, we first design a content-degradation feature disentanglement strategy to capture and isolate task-irrelevant features contained in the degraded image. Then, to avoid the catastrophic forgetting problem, we introduce a memory replay algorithm to further consolidate invariance knowledge learned from the previous pre-training to improve subsequent identity feature learning. In this way, our framework is able to continuously maintain degradation-invariant priors from one or more datasets to improve the robustness of identity features, achieving state-of-the-art Re-ID performance on several challenging real-world benchmarks with a unified model. Furthermore, the proposed framework can be extended to low-level image processing, e.g., low-light image enhancement, demonstrating the potential of our method as a general framework for the various vision tasks. Code and trained models will be available at: https://github.com/hyk1996/Degradation-Invariant-Re-ID-pytorch.

     

Robust Path Planning for Non-Holonomic Robots

This paper deals with mobile robots path planning. We decompose the problem in three parts. In the first part, we describe a modeling method based on a configuration space discretization. Each model element is built following a particular structure which is easy to handle, as we will show. We describe the methodologies and the algorithms allowing to build the model. In the second part, we propose a path-planning application for a non-holonomic robot in configuration space. In the third part, we modify the path in order to be robust according to the control errors.     

Learning from Innate Behaviors: A Quantitative Evaluation of Neural Network Controllers

The aim was to investigate a method of developing mobile robot controllers based on ideas about how plastic neural systems adapt to their environment by extracting regularities from the amalgamated behavior of inflexible (nonplastic) innate subsystems interacting with the world. Incremental bootstrapping of neural network controllers was examined. The objective was twofold. First, to develop and evaluate the use of prewired or innate robot controllers to bootstrap backpropagation learning for Multilayer Perceptron (MLP) controllers. Second, to develop and evaluate a new MLP controller trained on the back of another bootstrapped controller. The experimental hypothesis was that MLPs would improve on the performance of controllers used to train them. The performances of the innate and bootstrapped MLP controllers were compared in eight experiments on the tasks of avoiding obstacles and finding goals. Four quantitative measures were employed: the number of sensorimotor loops required to complete a task; the distance traveled; the mean distance from walls and obstacles; the smoothness of travel. The overall pattern of results from statistical analyses of these quantities supported the hypothesis; the MLP controllers completed the tasks faster, smoother, and steered further from obstacles and walls than their innate teachers. In particular, a single MLP controller incrementally bootstrapped by a MLP subsumption controller was superior to the others.     

Learning from Innate Behaviors: A Quantitative Evaluation of Neural Network Controllers

The aim was to investigate a method of developing mobile robot controllers based on ideas about how plastic neural systems adapt to their environment by extracting regularities from the amalgamated behavior of inflexible (non-plastic) innate s ubsystems interacting with the world.Incremental bootstrapping of neural network controllers was examined. The objective was twofold. First, to develop and evaluate the use of prewired or innate robot controllers to bootstrap backpropagation learning for Multi-Layer Perceptron (MLP) controllers. Second, to develop and evaluate a new MLP controller trained on the back of another bootstrapped controller. The experimental hypothesis was that MLPs would improve on the performance of controllers used to train them. The performances of the innate and bootstrapped MLP controllers were compared in eight experiments on the tasks of avoiding obstacles and finding goals. Four quantitative measures were employed: the number of sensorimotor loops required to complete a task; the distance traveled; the mean distance from walls and obstacles; the smoothness of travel. The overall pattern of results from statistical analyses of these quantities su pported the hypothesis; the MLP controllers completed the tasks faster, smoother, and steered further from obstacles and walls than their innate teachers. In particular, a single MLP controller incrementally bootstrapped by a MLP subsumption controller was superior to the others.     

针对动态非平衡数据集鲁棒的在线极端学习机

动态数据存在数据量动态改变,数据类别分布非平衡、不稳定等问题,这些问题成为分类的难点。针对该问题,通过对在线极端学习机模型进行拓展,提出鲁棒的权值在线极端学习机算法。为解决动态数据非平衡性,该算法借助代价敏感学习理论生成局部动态权值矩阵,从而优化分类模型产生的经验风险。同时,算法进一步考虑动态数据由于时序性质改变造成的数据分布变化,而引入遗忘因子增强分类器对数据分布变更的敏感性。算法在不同数据分布的24个非平衡动态数据集上测试,取得了较好的效果。     

空间机器人地面试验系统建模与规划

在空间机器人研究中,空间机器人的地面试验是必不可少的,通过地面试验系统模拟空间环境,可以进行实验,验证空间系统上所采用的理论、方法是否合理,为下一步的空间实验提供依据.为了给下一步的研究提供平台,针对小型智能空间机器人系统的气浮地面实验系统采用拉格朗日法建立系统的运动学、动力学模型,采用5-3-5混合插值法进行了路径规划,通过仿真试验,验证了空间建模理论和规划方法的正确性,为空间机器人实验的改进提供了依据.     

基于学习的智能机器人动作规划

本文回顾了基于学习的智能机器人动作规划:原理与系统的发展状况,评述了相应的基本原理和系统实现。     

Cooperative Mobile Robotics: Antecedents and Directions

There has been increased research interest in systems composed of multiple autonomous mobile robots exhibiting cooperative behavior. Groups of mobile robots are constructed, with an aim to studying such issues as group architecture, resource conflict, origin of cooperation, learning, and geometric problems. As yet, few applications of cooperative robotics have been reported, and supporting theory is still in its formative stages. In this paper, we give a critical survey of existing works and discuss open problems in this field, emphasizing the various theoretical issues that arise in the study of cooperative robotics. We describe the intellectual heritages that have guided early research, as well as possible additions to the set of existing motivations.     

从规划解中学习一阶派生谓词规则

派生谓词是描述动作非直接效果的主要方式.但是由人类专家设计的派生谓词规则(即领域理论)不能保证总是正确或者完备的,因此有时很难解释一个观察到的规划解为什么是有效的.结合归纳学习与分析学习的优点,文中提出一种称为FODRL(First-Order Derived Rules Learning)的算法,在不完美的初始领域理论的引导下从观察到的规划解中学习一阶派生谓词规则.FODRL基于归纳学习算法FOIL(First-Order Inductive Learning),最主要的改进是可以使用派生谓词的激活集来扩大搜索步,从而提高学习到的规则的精确度.学习过程分为两个步骤:先从规划解中提取训练例,然后学习能够最好拟合训练例和初始领域理论的一阶规则集.在PSR和PROME-LA两个派生规划领域进行实验,结果表明,在大部分情况下FODRL比FOIL(甚至包括其变型算法FOCL)学习到的规则的精确度都要高.     

一种鲁棒性的少样本学习方法

少样本学习是目前机器学习研究领域的一个热点,它能在少量的标记样本中学习到较好的分类模型.但是,在噪声的不确定环境中,传统的少样本学习模型泛化能力弱.针对这一问题,提出一种鲁棒性的少样本学习方法RFSL(Robust Few-Shot Learning).首先,使用核密度估计(Kernel Density Estimat...     

Robotics explorations to enhance information technology literacy in rural schools

This paper describes a project in ＂Biomechanics and Robotics Explorations for Information Technology Literacy and Skills in Rular Schools＂ at the Department of Engineering at East Carolina University funded by the National Science Foundation. The project is designed to show how mathematics and science teachers can combine information technology explorations with mathematical problem solving to develop context rich lessons that attract rural students to careers in science, engineering and technology. Over the summer of 2007, 30 teachers and 60 students participated in a camp designed to provide both theoretical and hands-on experiences in various STEM facets. The teachers spent two weeks at East Carolina University attending workshops in various areas including robotics, biomechanics, solid modeling, and other IT literacy areas in preparation for integration into Math and Science curricula in their schools. Selected students from the same schools spent three weeks going through similar workshops. This paper will focus on the robotics component of the project and its implementation at the Department of Engineering at East Carolina University. The paper presents information about robotics summer academies for teachers and for students, as well as our experiences from the academies. In addition, information from surveys that were administered in both teacher and student workshops are presented.     

Rehabilitation Robotics: Performance-Based Progressive Robot-Assisted Therapy

In this paper we describe the novel concept of performance-based progressive robot therapy that uses speed, time, or EMG thresholds to initiate robot assistance. We pioneered the clinical application of robot-assisted therapy focusing on stroke—the largest cause of disability in the US. We have completed several clinical studies involving well over 200 stroke patients. Research to date has shown that repetitive task-specific, goal-directed, robot-assisted therapy is effective in reducing motor impairments in the affected arm after stroke. One research goal is to determine the optimal therapy tailored to each stroke patient that will maximize his/her recovery. A proposed method to achieve this goal is a novel performance-based impedance control algorithm, which is triggered via speed, time, or EMG. While it is too early to determine the effectiveness of the algorithm, therapists have already noted one very strong benefit, a significant reduction in arm tone.     

Learning to Understand Information on the Internet: An Example-Based Approach

The explosive growth of the Web has made intelligent softwareassistants increasingly necessary for ordinary computer users. Bothtraditional approaches—search engines, hierarchical indices—andintelligent software agents require significant amounts of humaneffort to keep up with the Web. As an alternative, we investigate theproblem of automatically learning to interact with informationsources on the Internet. We report on ShopBotand ILA , two implemented agents that learn touse such resources. ShopBot learns how to extract information from onlinevendors using only minimal knowledge about product domains. Giventhe home pages of several online stores, ShopBotautonomously learns how to shop at those vendors. After its learningis complete, ShopBot is able to speedily visitover a dozen software stores and CD vendors, extract productinformation, and summarize the results for the user. ILAlearns to translate information from Internetsources into its own internal concepts. ILAbuilds a model of an information source that specifies the translation between the source's output and ILA 's model of the world. ILA iscapable of leveraging a small amount of knowledge about a domain tolearn models of many information sources. We show that ILA 's learning is fast and accurate, requiring only a smallnumber of queries per information source.     

Reinforcement Learning Agents

Reinforcement Learning (RL) is learning through directexperimentation. It does not assume the existence of a teacher thatprovides examples upon which learning of a task takes place. Instead, inRL experience is the only teacher. With historical roots on the study ofbiological conditioned reflexes, RL attracts the interest of Engineersand Computer Scientists because of its theoretical relevance andpotential applications in fields as diverse as Operational Research andIntelligent Robotics.Computationally, RL is intended to operate in a learning environmentcomposed by two subjects: the learner and a dynamic process. Atsuccessive time steps, the learner makes an observation of the processstate, selects an action and applies it back to the process. Its goal isto find out an action policy that controls the behavior of the dynamicprocess, guided by signals (reinforcements) that indicate how badly orwell it has been performing the required task. These signals are usuallyassociated to a dramatic condition – e.g., accomplishment of a subtask(reward) or complete failure (punishment), and the learner tries tooptimize its behavior by using a performance measure (a function of thereceived reinforcements). The crucial point is that in order to do that,the learner must evaluate the conditions (associations between observedstates and chosen actions) that led to rewards or punishments.Starting from basic concepts, this tutorial presents the many flavorsof RL algorithms, develops the corresponding mathematical tools, assesstheir practical limitations and discusses alternatives that have beenproposed for applying RL to realistic tasks.     

Multi-Layer Hierarchical Rule Learning in Reactive Robot Control Using Incremental Decision Trees

This paper presents a new approach to the intelligent navigation of a mobile robot. The hybrid control architecture described combines properties of purely reactive and behaviour-based systems, providing the ability both to learn automatically behaviours from inception, and to capture these in a distributed hierarchy of decision tree networks. The robot is first trained in the simplest world which has no obstacles, and is then trained in successively more complex worlds, using the knowledge acquired in the previous worlds. Each world representing the perceptual space is thus directly mapped on a unique rule layer which represents in turn the robot action space encoded in a distinct decision tree. A major advantage of the current implementation, compared with the previous work, is that the generated rules are easily understood by human users. The paper demonstrates that the proposed behavioural decomposition approach provides efficient management of complex knowledge, and that the learning mechanism is able to cope with noise and uncertainty in sensory data.