Managing Complexity in Large Learning Robotic Systems

Autonomous learning systems of significant complexity often consist of several interacting modules or agents. These modules collaborate to produce a system which, when viewed as a whole, exhibit behaviour that can be interpreted in some way as learning. In designing these systems, the complexity of the interactions of large numbers of modules can become overwhelming, making debugging difficult and obscuring the workings of the system when viewed from an architectural level. A way of controlling system complexity called the Layered Learning System architecture (LLS) has been developed, which offers the advantages of incremental development and testing, easier debugging and progressive upgrading facilitation. A hexapod robot has been implemented using LLS principles, with the main learning task being that of learning to walk as fast as possible without falling over.     

Scheduling in Reentrant Robotic Cells: Algorithms and Complexity

We study the scheduling of m-machine reentrant robotic cells, in which parts need to reenter machines several times before they are finished. The problem is to find the sequence of 1-unit robot move cycles and the part processing sequence which jointly minimize the cycle time or the makespan. When m = 2, we show that both the cycle time and the makespan minimization problems are polynomially solvable. When m = 3, we examine a special class of reentrant robotic cells with the cycle time objective. We show that in a three-machine loop-reentrant robotic cell, the part sequencing problem under three out of the four possible robot move cycles for producing one unit is strongly -hard. The part sequencing problem under the remaining robot move cycle can be solved easily. Finally, we prove that the general problem, without restriction to any robot move cycle, is also intractable.     

关于TSP的计算复杂性

本文研究TSP的计算复杂性,指出了判定型TSP与最优型TSP具有不同的计算复杂性,同时指出了文献[1]-[3]中的一些错误。     

A Robotic Excavator for Autonomous Truck Loading

Excavators are used for the rapid removal of soil and other materials in mines, quarries, and construction sites. The automation of these machines offers promise for increasing productivity and improving safety. To date, most research in this area has focussed on selected parts of the problem. In this paper, we present a system that completely automates the truck loading task. The excavator uses two scanning laser rangefinders to recognize and localize the truck, measure the soil face, and detect obstacles. The excavator's software decides where to dig in the soil, where to dump in the truck, and how to quickly move between these points while detecting and stopping for obstacles. The system was fully implemented and was demonstrated to load trucks as fast as human operators.     

机器人操作臂跟踪动态目标的轨迹规划方法

针对机器人操作臂跟踪运动目标问题,提出一种基于遗传算法的轨迹规划方法。通过对关节加速度的增量进行编码,实现在操作臂的关节空间进行轨迹优化,得到操作臂在跟踪运动目标过程中所需要的轨迹。仿真计算的结果表明,所提出的方法是有效的。     

Multi-Modal Locomotion Robotic Platform Using Leg-Track-Wheel Articulations

Other than from its sensing and processing capabilities, a mobile robotic platform can be limited in its use by its ability to move in the environment. Legs, tracks and wheels are all efficient means of ground locomotion that are most suitable in different situations. Legs allow to climb over obstacles and change the height of the robot, modifying its viewpoint of the world. Tracks are efficient on uneven terrains or on soft surfaces (snow, mud, etc.), while wheels are optimal on flat surfaces. Our objective is to work on a new concept capable of combining different locomotion mechanisms to increase the locomotion capabilities of the robotic platform. The design we came up with, called AZIMUT, is symmetrical and is made of four independent leg-track-wheel articulations. It can move with its articulations up, down or straight, allowing the robot to deal with three-dimensional environments. AZIMUT is also capable of moving sideways without changing its orientation, making it omnidirectional. By putting sensors on these articulations, the robot can also actively perceive its environment by changing the orientation of its articulations. Designing a robot with such capabilities requires addressing difficult design compromises, with measurable impacts seen only after integrating all of the components together. Modularity at the structural, hardware and embedded software levels, all considered concurrently in an iterative design process, reveals to be key in the design of sophisticated mobile robotic platforms.This research is supported financially by the Canada Research Chair (CRC) program, the Natural Sciences and Engineering Research Council of Canada (NSERC), the Canadian Foundation for Innovation (CFI) and the Faculty of Engineering of the Université de Sherbrooke. Patent pending.François Michaud is the Canada Research Chairholder in Autonomous Mobile Robots and Intelligent Systems, and an Associate Professor at the Department of Electrical Engineering and Computer Engineering of the Université de Sherbrooke. He is the principal investigator of LABORIUS, a research laboratory working on applying AI methodologies in the design of intelligent autonomous systems that can assist humans in everyday lives. His research interests are architectural methodologies for intelligent decision making, autonomous mobile robotics, social robotics, robot learning and intelligent systems. He received his bachelors degree, Masters degree and Ph.D. degree in Electrical Engineering from the Université de Sherbrooke. He is a member of IEEE, AAAI and OIQ (Ordre des ingénieurs du Québec).Dominic Létourneau has a Bachelor degree in Computer Engineering and a Masters degree in Electrical Engineering from the Université de Sherbrooke. Since 2001, he is a research engineer at the LABORIUS Mobile Robotics and Intelligent Systems Laboratory. His research interests cover combination of systems and intelligent capabilities to increase the usability of autonomous mobile robots in the real world. His expertise lies in artificial vision, mobile robotics, robot programming and integrated design. He is a member of OIQ (Ordre des ingénieurs du Québec).Martin Arsenault has a Bachelor degree in Mechanical Engineering and from the Université de Sherbrooke. For AZIMUT, he developed the Direction subsystem.Yann Bergeron is a Mechanical Junior Engineer working in a consulting engineering firm, Groupe HBA. His field of activity is most oriented in industrial projects and construction. He received his bachelors degree in Mechanical Engineering from the University de Sherbrooke. For AZIMUT, he worked on the Track-Wheel subsystem. He is a member of OIQ (Ordre des ingénieurs du Québec) and JECQ (Jeunes entrepreneurs du Centre-du-Québec).Richard Cadrin has a Bachelor degree in Mechanical Engineering from the Université de Sherbrooke. For AZIMUT, he worked on the Propulstion subsystem.Frédéric Gagnon has a Bachelor degree in Mechanical Engineering and from the Université de Sherbrooke. For AZIMUT, he designed the chassis and worked on the integration of the articulations, the bodywork and the hardware. He is currently a research engineer at LABORIUS, working on AZIMUTs second prototype and other robotic projects. He also contributes in a mobile robotic.Marc-Antoine Legault has a Bachelor degree in Mechanical Engineering from the Université de Sherbrooke, and is currently pursuing a Masters degree at LABORIUS on serial-elastic actuators. He worked on AZIMUTs Propulsion subsystem. He also works on the design of other mobile robotic platforms. He is a member of OIQ (Ordre des ingénieurs du Québec).Mathieu Millette has a Bachelor degree in Mechanical Engineering from the Université de Sherbrooke. For AZIMUT, he developed the Tensor subsystem and he designed and integrated the battery support inside the chassis. He is now working as a junior mechanical engineer in process and technical development in a third sector mill.Jean-François Paré graduated in Mechanical Engineering from the Université de Sherbrooke. He is also trained as a professional coach from R.I.C.K. (Réseau International de Coaching Kokopelli). He launched is own business in individual coaching, team coaching, team building and coaching workshops. For businesses and individuals, he assists people in reaching their well-being and efficiency and to manifest their leadership.Marie-Christine Tremblay has obtained a Bachelor degree in Mechanical Engineering from the Université de Sherbrooke. As a part of the AZIMUTõs team, she contributed the Track-Wheel subsystem and developed a new track with high adherence and wear resistance to climb stairs. She is working in the Engineering Department at Hydra-Fab Industriel, a company conceiving electro-hydraulic systems for high speed trains.Pierre Lepage has a Computer Engineering degree from the Université de Sherbrooke. For AZIMUT, he designed the user interface to control the robot and the propulsion system. He is currently a research engineer at LABORIUS, working on AZIMUTs second prototype and other robotic projects. He is involved in research projects on coordinated behavior of a group of mobile robots. He is also actively involved in a mobile robotic startup.Yan Morin has an Electrical Engineering degree from the Université de Sherbrooke. For AZIMUT, he designed the electrical interface to control the motors of the robot. He is currently a research engineer at LABORIUS, working on AZIMUTs second prototype and other robotic projects. He is also actively involved in a mobile robotic startup.Jonathan Bisson has a Bachelor degree in Computer Engineering and a Masters degree in Electrical Engineering from the Université de Sherbrooke. His contribution to AZIMUT was on the modular distributed computing architecture. His expertise lies in electronics, motor control, embedded systems and ultrasonic transducers.Serge Caron is technician in computer systems at the Department of Electrical Engineering and Computer Engineering of the Universit de Sherbrooke. His interests are in designing mobile robotic platforms, from four-legged to wheeled robots. He is also a writer for hobbyist robotic journals.     

仿生机器鱼的检测与特征提取

在现有仿生机器鱼平台的基础上研究了仿生机器鱼的检测和特征提取算法,并采用均值背景模型结合大津法计算的阈值计算出机器鱼位置,采用前后背景的自适应更新方式减少光照改变的影响以及机器鱼影子的干扰。然后通过数学形态学处理,得到最终的二值图像并将符合要求的二值图像块数目标记在左上角。将目标块用外接矩形围住,通过相应的公式计算出各机器鱼二值图像的几何特征。实验结果证明可以达到预期要求。     

机械手自适应模糊控制器设计

本文提出了一种自适应模糊控制器用于机械手轨迹跟踪控制,该控制器通过对训练数据的聚类分析提取典型数据送入神经网络学习,得到的控制规则更加适用,在线控制时可以自动调整神经网络的结构以及隶属函数的参数。仿真实验表明该控制器能够对机械手实时控制,而且精度较高。     

Sensory-Based Colour Sorting Automated Robotic Cell

Robotics application in colour recognition using fiber optic cabled sensors interfaced with robot controller and Programmable Logic Controller (PLC) is discussed in this paper. The sensors send input signals to the robot controller and the specified program will be executed with respect to the triggered input. The aim of this research work is to recognize colour by pin point detection and sorting of object specimens with respect to their colour attributes, which includes hue, saturation and luminance level. The controller programs were designed to control the robot and the conveyor belt independently parallel to each other via relays, to be synchronized during operation. Finally, the calculative results were verified experimentally and the real time implementation was carried out. It can be observed how controllers are integrated and synchronized in a system to perform a desired operation without conflict using real time applications such as chemical, pharmaceutical, agricultural, food industries and even recycling.     

A Control Module Scheme for an Underactuated Underwater Robotic Vehicle

Despite major advances in Autonomous Underwater Vehicle design, the manually operated Underwater Vehicle (ROV) is still very much the industry workhorse. Current technologies are being used to reduce the stress of direct task operations by providing autonomy and to improve efficiency. This paper presents a design of a control module subsystem for a VE tele-operated ROV system. It discusses the design and implementation of the control module. Using modelling, simulation and experiments, the vehicle model and its parameters have been identified. These are used in the analysis and design of closed loop stabilising controllers for station keeping. As the vehicle has fewer actuators than possible degrees of freedom, it is necessary to limit the controllable degrees of freedom. These variables are eventually selected based on the inherent vehicle dynamics. Using the Lyapunov direct method, appropriate stabilising controllers have been designed. The station-keeping mode controller has PID structure and its gain values are designed using a non-linear optimising approach. Simulation and swimming pool tests for the heave and yaw directions have shown that the control module is able to provide reasonable depth and heading station keeping.     

Optimal and Robust Control of Multi DOF Robotic Manipulator: Design and Hardware Realization

Robots have become an integral part of industrial automation. Their ultimate role and contribution in this sector is essentially a function of the associated control strategy to ensure precision, repeatability, and reliability, particularly in an environment polluted with disturbances and uncertainties. This research aims to present a design of the modern control strategies for a 6 degree of freedom robotic manipulator. Based on derived kinematic and dynamic models of the robot, optimal and robust control strategies are simulated and practically realized on a custom developed pseudo-industrial framework named as AUTonomous Articulated Robotic Educational Platform. Results of the experimental trials in terms of trajectory tracking demonstrate efficiency and usefulness of the presented control approaches.     

A Constraint-Based Robotic Soccer Team

It is a challenging task for a team of multiple fast-moving robots to cooperate with each other and to compete with another team in a dynamic, real-time environment. For a robot team to play soccer successfully, various technologies have to be incorporated including robotic architecture, multi-agent collaboration and real-time reasoning. A robot is an integrated system, with a controller embedded in its plant. A robotic system is the coupling of a robot to its environment. Robotic systems are, in general, hybrid dynamic systems, consisting of continuous, discrete and event-driven components. Constraint Nets (CN) provide a semantic model for modeling hybrid dynamic systems. Controllers are embedded constraint solvers that solve constraints in real-time. A controller for our robot soccer team, UBC Dynamo98, has been modeled in CN, and implemented in Java, using the Java Beans architecture. A coach program using an evolutionary algorithm has also been designed and implemented to adjust the weights of the constraints and other parameters in the controller. The results demonstrate that the formal CN approach is a practical tool for designing and implementing controllers for robots in multi-agent real-time environments. They also demonstrate the effectiveness of applying the evolutionary algorithm to the CN-modeled controllers.     

Integration of Heterogeneity for Human-Friendly Robotic Operations

The objective of this paper is to develop an analytical scheme to integrate the heterogeneity of human and robot functions to achieve a human-friendly robotic operations. The heterogeneity of human and robot functions can be characterized by the fact that humans are intelligent while robots are fast, powerful and accurate. Humans can use their knowledge and experience to quickly respond to unexpected events, which makes it easy for humans to deal with unstructured environments. In contrast, robots can easily enhance the mechanical power of humans and the ability of humans to work remotely. Therefore, robots are capable of performing precise and repetitive tasks at high speed or in a hazardous environment. The important issue, in light of human/robot heterogeneity, is how to plan and control a robotic operation such that the human and the robot can cooperate in a complementary manner. Thus, a task which cannot be done by either human or robot alone can be performed efficiently and robustly by both. This paper introduces a new paradigm for human/robot interactive systems, heterogeneous function-based human/robot cooperation. A new perceptive action reference frame has been developed in the paper. It matches human perception and robot sensory measurement, and provides a platform for modeling the human/robot cooperative operations. The theoretical results presented in the paper have laid down a foundation for stability analysis as well as a planning and control system design of human/robot integrated systems. The implementations and experimental results have clearly demonstrated the advantages of proposed methods.     

用于汽车耐久性试验的驾驶机器人模糊控制仿真

介绍一种用于汽车耐久性试验驾驶机器人模糊控制方法。驾驶机器人由交流伺服电机、控制器、滚珠丝杠以及机械定位装置组成。阐述了模糊控制的原理、模糊PID控制器的设计规则和设计过程。在MATLAB环境仿真模拟机器人位置伺服系统,表明了模糊PID控制器的有效性。     

机器人机械手臂关节驱动控制系统设计

基于动态方程和Hamilton-Jacobi原理,从硬件到软件,设计了机器人机械手臂的关节驱动控制系统。主要任务是使该系统的能耗最小化。仿真和实验的结果,证明了该系统的可用性和准确性。     

Experience-Based Iterative Learning Controllers for Robotic Systems

An experience based iterative learning controller is proposed for a general class of robotic systems. Experience of the iterative learning controller is stored in the memory in terms of input output data and later used for the prediction of the initial control input for a new desired trajectory. It is proved in this paper that using this approach we can reduce the number of iterations to achieve a certain user defined tracking accuracy. This approach is very general and applicable to all kinds of existing iterative learning control schemes. Numerical illustrations showed the effectiveness of the proposed method.     

Position and Force Control of Two Constrained Robotic Manipulators

In this paper, the problem of controlling two robotic manipulators handling a constrained object is addressed. First, a reduced order dynamic model of the system is obtained. Using this model, a controller that guarantees the asymptotic convergence of the position, the internal force, and the constraint force to their desired values is proposed. Simulation results for two three-link planar manipulators moving a constrained object demonstrate the effectiveness of the proposed controller.     

仿生蝠鲼机器鱼BH-RAY3的研制及水力实验

通过对蝠鲼外形参数的设计和胸鳍变形机理的研究,研制了仿生蝠鲼机器鱼BH-RAY3。该机器鱼具有双柔性胸鳍,翼展620 mm,身长400 mm,最大航速1.09 DLS,外形阻力小,通过尾舵实现升降及转向。对BH-RAY3进行的水洞拖拽实验表明,随着拍动频率和幅度的增大,胸鳍推进力增大,在频率为1.0 Hz,幅度为24°时达到1.452 N;其次,对胸鳍被动变形形态的分析表明,柔性胸鳍同时产生0.25波长的弦向波和1个波长的展向波,与蝠鲼短波长、低频率的运动特点一致,验证了设计的合理性和可行性。     

Three-Dimensional Robotic Vision Using Ultrasonic Sensors

This article describes a three-dimensional artificial vision system for robotic applications using an ultrasonic sensor array. The array is placed on the robot grip so that it is possible to detect the presence of an object, to direct the robot tool towards it, and to locate the object position. It will provide visual information about the object's surface by means of superficial scanning and it permits the object shape reconstruction. The developed system uses an approximation of the ultrasonic radiation and reception beam shape for calculating the first contact points with the object's surface. On the other hand, the position of the array's sensors has been selected in order to provide the sensorial head with other useful capabilities, such as edge detection and edge tracking. Furthermore, the article shows the structure of the sensorial head for avoiding successive rebounds between the head and the object surface, and for eliminating the mechanical vibrations among sensors.     

Sequencing and Scheduling in Robotic Cells: Recent Developments

A great deal of work has been done to analyze the problem of robot move sequencing and part scheduling in robotic flowshop cells. We examine the recent developments in this literature. A robotic flowshop cell consists of a number of processing stages served by one or more robots. Each stage has one or more machines that perform that stage’s processing. Types of robotic cells are differentiated from one another by certain characteristics, including robot type, robot travel-time, number of robots, types of parts processed, and use of parallel machines within stages. We focus on cyclic production of parts. A cycle is specified by a repeatable sequence of robot moves designed to transfer a set of parts between the machines for their processing.We start by providing a classification scheme for robotic cell scheduling problems that is based on three characteristics: machine environment, processing restrictions, and objective function, and discuss the influence of these characteristics on the methods of analysis employed. In addition to reporting recent results on classical robotic cell scheduling problems, we include results on robotic cells with advanced features such as dual gripper robots, parallel machines, and multiple robots. Next, we examine implementation issues that have been addressed in the practice-oriented literature and detail the optimal policies to use under various combinations of conditions. We conclude by describing some important open problems in the field.