View Chapter

Chapter 15 — Robot Learning

Jan Peters, Daniel D. Lee, Jens Kober, Duy Nguyen-Tuong, J. Andrew Bagnell and Stefan Schaal

Machine learning offers to robotics a framework and set of tools for the design of sophisticated and hard-to-engineer behaviors; conversely, the challenges of robotic problems provide both inspiration, impact, and validation for developments in robot learning. The relationship between disciplines has sufficient promise to be likened to that between physics and mathematics. In this chapter, we attempt to strengthen the links between the two research communities by providing a survey of work in robot learning for learning control and behavior generation in robots. We highlight both key challenges in robot learning as well as notable successes. We discuss how contributions tamed the complexity of the domain and study the role of algorithms, representations, and prior knowledge in achieving these successes. As a result, a particular focus of our chapter lies on model learning for control and robot reinforcement learning. We demonstrate how machine learning approaches may be profitably applied, and we note throughout open questions and the tremendous potential for future research.

Inverse reinforcement

Author  Pieter Abbeel

Video ID : 353

This video shows a successful example of inverse reinforcement learning for acrobatic helicopter maneuvers. It illustrates apprenticeship learning algorithms, which leverage expert demonstrations to efficiently learn good controllers for tasks as demonstrated by an expert. The experimental results captured here include the first autonomous execution of a wide range of maneuvers and a complete airshow. The controllers perform as well as, and often even better than, the human expert pilot. The video illustrates a solution to the "Curse of Goal Specification" in Sect 15.3.6 Challenges in Robot Reinforcement Learning. Reference: P. Abbeel, A. Coates, A.Y. Ng: Autonomous helicopter aerobatics through apprenticeship learning, Int. J. Robot. Res. 29(13), 1608–1639 (2010)

Chapter 61 — Robot Surveillance and Security

Wendell H. Chun and Nikolaos Papanikolopoulos

This chapter introduces the foundation for surveillance and security robots for multiple military and civilian applications. The key environmental domains are mobile robots for ground, aerial, surface water, and underwater applications. Surveillance literallymeans to watch fromabove,while surveillance robots are used to monitor the behavior, activities, and other changing information that are gathered for the general purpose of managing, directing, or protecting one’s assets or position. In a practical sense, the term surveillance is taken to mean the act of observation from a distance, and security robots are commonly used to protect and safeguard a location, some valuable assets, or personal against danger, damage, loss, and crime. Surveillance is a proactive operation,while security robots are a defensive operation. The construction of each type of robot is similar in nature with amobility component, sensor payload, communication system, and an operator control station.

After introducing the major robot components, this chapter focuses on the various applications. More specifically, Sect. 61.3 discusses the enabling technologies of mobile robot navigation, various payload sensors used for surveillance or security applications, target detection and tracking algorithms, and the operator’s robot control console for human–machine interface (HMI). Section 61.4 presents selected research activities relevant to surveillance and security, including automatic data processing of the payload sensors, automaticmonitoring of human activities, facial recognition, and collaborative automatic target recognition (ATR). Finally, Sect. 61.5 discusses future directions in robot surveillance and security, giving some conclusions and followed by references.

Indoor, urban aerial vehicle navigation

Author  Jonathan How

Video ID : 703

The MIT indoor multi-vehicle testbed is specially designed to study long duration missions in a controlled, urban environment. This testbed is being used to implement and analyze the performance of techniques for embedding the fleet and vehicle health state into the mission and UAV planning. More than four air vehicles can be flown in a typical-sized room, and it takes no more than one operator to set up the platform for flight testing at any time of day and for any length of time. At the heart of the testbed is a global metrology system that yields very accurate, high bandwidth position and attitude data for all vehicles in the entire room.

Chapter 65 — Domestic Robotics

Erwin Prassler, Mario E. Munich, Paolo Pirjanian and Kazuhiro Kosuge

When the first edition of this book was published domestic robots were spoken of as a dream that was slowly becoming reality. At that time, in 2008, we looked back on more than twenty years of research and development in domestic robotics, especially in cleaning robotics. Although everybody expected cleaning to be the killer app for domestic robotics in the first half of these twenty years nothing big really happened. About ten years before the first edition of this book appeared, all of a sudden things started moving. Several small, but also some larger enterprises announced that they would soon launch domestic cleaning robots. The robotics community was anxiously awaiting these first cleaning robots and so were consumers. The big burst, however, was yet to come. The price tag of those cleaning robots was far beyond what people were willing to pay for a vacuum cleaner. It took another four years until, in 2002, a small and inexpensive device, which was not even called a cleaning robot, brought the first breakthrough: Roomba. Sales of the Roomba quickly passed the first million robots and increased rapidly. While for the first years after Roomba’s release, the big players remained on the sidelines, possibly to revise their own designs and, in particular their business models and price tags, some other small players followed quickly and came out with their own products. We reported about theses devices and their creators in the first edition. Since then the momentum in the field of domestics robotics has steadily increased. Nowadays most big appliance manufacturers have domestic cleaning robots in their portfolio. We are not only seeing more and more domestic cleaning robots and lawn mowers on the market, but we are also seeing new types of domestic robots, window cleaners, plant watering robots, tele-presence robots, domestic surveillance robots, and robotic sports devices. Some of these new types of domestic robots are still prototypes or concept studies. Others have already crossed the threshold to becoming commercial products.

For the second edition of this chapter, we have decided to not only enumerate the devices that have emerged and survived in the past five years, but also to take a look back at how it all began, contrasting this retrospection with the burst of progress in the past five years in domestic cleaning robotics. We will not describe and discuss in detail every single cleaning robot that has seen the light of the day, but select those that are representative for the evolution of the technology as well as the market. We will also reserve some space for new types of mobile domestic robots, which will be the success stories or failures for the next edition of this chapter. Further we will look into nonmobile domestic robots, also called smart appliances, and examine their fate. Last but not least, we will look at the recent developments in the area of intelligent homes that surround and, at times, also control the mobile domestic robots and smart appliances described in the preceding sections.

Husqvarna Automower vs competitors

Author  Erwin Prassler

Video ID : 731

Video shows a comparison of the Automower of Husquarna with the products of competitors such as Friendly Machines, John Deer, and Honda.

Chapter 11 — Robots with Flexible Elements

Alessandro De Luca and Wayne J. Book

Design issues, dynamic modeling, trajectory planning, and feedback control problems are presented for robot manipulators having components with mechanical flexibility, either concentrated at the joints or distributed along the links. The chapter is divided accordingly into two main parts. Similarities or differences between the two types of flexibility are pointed out wherever appropriate.

For robots with flexible joints, the dynamic model is derived in detail by following a Lagrangian approach and possible simplified versions are discussed. The problem of computing the nominal torques that produce a desired robot motion is then solved. Regulation and trajectory tracking tasks are addressed by means of linear and nonlinear feedback control designs.

For robots with flexible links, relevant factors that lead to the consideration of distributed flexibility are analyzed. Dynamic models are presented, based on the treatment of flexibility through lumped elements, transfer matrices, or assumed modes. Several specific issues are then highlighted, including the selection of sensors, the model order used for control design, and the generation of effective commands that reduce or eliminate residual vibrations in rest-to-rest maneuvers. Feedback control alternatives are finally discussed.

In each of the two parts of this chapter, a section is devoted to the illustration of the original references and to further readings on the subject.

State feedback response to impulse in presence of link flexibility

Author  Wayne Book

Video ID : 781

A laboratory gantry robot with a final flexible link is excited by an external impulse disturbance. The video shows the effective damping obtained using full state feedback control with an accurately tuned estimator. The reduction in settling time compared to PID joint control is dramatic. This is one of two coordinated videos, the other showing the same experiment under PID control. Reference: B. Post: Robust State Estimation for the Control of Flexible Robotic Manipulators, Dissertation, School of Mechanical Engineering, Georgia Institute of Technology, Atlanta (2013)

Chapter 47 — Motion Planning and Obstacle Avoidance

Javier Minguez, Florant Lamiraux and Jean-Paul Laumond

This chapter describes motion planning and obstacle avoidance for mobile robots. We will see how the two areas do not share the same modeling background. From the very beginning of motion planning, research has been dominated by computer sciences. Researchers aim at devising well-grounded algorithms with well-understood completeness and exactness properties.

The challenge of this chapter is to present both nonholonomic motion planning (Sects. 47.1–47.6) and obstacle avoidance (Sects. 47.7–47.10) issues. Section 47.11 reviews recent successful approaches that tend to embrace the whole problemofmotion planning and motion control. These approaches benefit from both nonholonomic motion planning and obstacle avoidance methods.

Autonomous robotic smart-wheelchair navigation in an urban environment

Author  VADERlab

Video ID : 707

This video demonstrates the reliable navigation of a smart wheelchair system (SWS) in an urban environment. Urban environments present unique challenges for service robots. They require localization accuracy at the sidewalk level, but compromise estimated GPS positions through significant multipath effects. However, they are also rich in landmarks that can be leveraged by feature-based localization approaches. To this end, the SWS employed a map-based approach. A map of South Bethlehem was acquired using a server vehicle, synthesized a priori, and made accessible to the SWS client. The map embedded not only the locations of landmarks, but also semantic data delineating seven different landmark classes to facilitate robust data association. Landmark segmentation and tracking by the SWS was then accomplished using both 2-D and 3-D LIDAR systems. The resulting localization algorithm has demonstrated decimeter-level positioning accuracy in a global coordinate frame. The localization package was integrated into a ROS framework with a sample-based planner and control loop running at 5 Hz. For validation, the SWS repeatedly navigated autonomously between Lehigh University's Packard Laboratory and the University bookstore, a distance of approximately 1.0 km roundtrip.

Chapter 18 — Parallel Mechanisms

Jean-Pierre Merlet, Clément Gosselin and Tian Huang

This chapter presents an introduction to the kinematics and dynamics of parallel mechanisms, also referred to as parallel robots. As opposed to classical serial manipulators, the kinematic architecture of parallel robots includes closed-loop kinematic chains. As a consequence, their analysis differs considerably from that of their serial counterparts. This chapter aims at presenting the fundamental formulations and techniques used in their analysis.

Par2 robot

Author  Sébastien Krut

Video ID : 51

This video demonstrates the Par2 robot, a high-speed planar parallel robot.

Chapter 20 — Snake-Like and Continuum Robots

Ian D. Walker, Howie Choset and Gregory S. Chirikjian

This chapter provides an overview of the state of the art of snake-like (backbones comprised of many small links) and continuum (continuous backbone) robots. The history of each of these classes of robot is reviewed, focusing on key hardware developments. A review of the existing theory and algorithms for kinematics for both types of robot is presented, followed by a summary ofmodeling of locomotion for snake-like and continuum mechanisms.

OctArms I-V

Author  Ian Walker

Video ID : 158

Video showing five different iterations of the OctArm continuum manipulator.

Chapter 59 — Robotics in Mining

Joshua A. Marshall, Adrian Bonchis, Eduardo Nebot and Steven Scheding

This chapter presents an overview of the state of the art in mining robotics, from surface to underground applications, and beyond. Mining is the practice of extracting resources for utilitarian purposes. Today, the international business of mining is a heavily mechanized industry that exploits the use of large diesel and electric equipment. These machines must operate in harsh, dynamic, and uncertain environments such as, for example, in the high arctic, in extreme desert climates, and in deep underground tunnel networks where it can be very hot and humid. Applications of robotics in mining are broad and include robotic dozing, excavation, and haulage, robotic mapping and surveying, as well as robotic drilling and explosives handling. This chapter describes how many of these applications involve unique technical challenges for field roboticists. However, there are compelling reasons to advance the discipline of mining robotics, which include not only a desire on the part of miners to improve productivity, safety, and lower costs, but also out of a need to meet product demands by accessing orebodies situated in increasingly challenging conditions.

Autonomous loading of fragmented rock

Author  Joshua Marshall

Video ID : 718

This video shows autonomous loading of fragmented rock, first on a 1-t capacity Kubota loader at Kingston, Canada, followed by an implementation on a 14-t capacity Atlas Copco ST14 LHD in an underground mine at Kvarntorp, Sweden. The algorithm used in these demonstrations is based on force-feedback sensed in the loader cylinder pressures and utilizes an admittance control structure.

Chapter 39 — Cooperative Manipulation

Fabrizio Caccavale and Masaru Uchiyama

This chapter is devoted to cooperative manipulation of a common object by means of two or more robotic arms. The chapter opens with a historical overview of the research on cooperativemanipulation, ranging from early 1970s to very recent years. Kinematics and dynamics of robotic arms cooperatively manipulating a tightly grasped rigid object are presented in depth. As for the kinematics and statics, the chosen approach is based on the socalled symmetric formulation; fundamentals of dynamics and reduced-order models for closed kinematic chains are discussed as well. A few special topics, such as the definition of geometrically meaningful cooperative task space variables, the problem of load distribution, and the definition of manipulability ellipsoids, are included to give the reader a complete picture ofmodeling and evaluation methodologies for cooperative manipulators. Then, the chapter presents the main strategies for controlling both the motion of the cooperative system and the interaction forces between the manipulators and the grasped object; in detail, fundamentals of hybrid force/position control, proportional–derivative (PD)-type force/position control schemes, feedback linearization techniques, and impedance control approaches are given. In the last section further reading on advanced topics related to control of cooperative robots is suggested; in detail, advanced nonlinear control strategies are briefly discussed (i. e., intelligent control approaches, synchronization control, decentralized control); also, fundamental results on modeling and control of cooperative systems possessing some degree of flexibility are briefly outlined.

Control of cooperative manipulators in the operational space

Author  Oussama Khatib

Video ID : 70

This video shows a series of experiments on operational space control of cooperative manipulators. Both the virtual linkage and augmented object concepts are experimentally demonstrated, together with cooperative manipulation via multiple mobile arms (Romeo & Juliet).

Chapter 67 — Humanoids

Paul Fitzpatrick, Kensuke Harada, Charles C. Kemp, Yoshio Matsumoto, Kazuhito Yokoi and Eiichi Yoshida

Humanoid robots selectively immitate aspects of human form and behavior. Humanoids come in a variety of shapes and sizes, from complete human-size legged robots to isolated robotic heads with human-like sensing and expression. This chapter highlights significant humanoid platforms and achievements, and discusses some of the underlying goals behind this area of robotics. Humanoids tend to require the integration ofmany of the methods covered in detail within other chapters of this handbook, so this chapter focuses on distinctive aspects of humanoid robotics with liberal cross-referencing.

This chapter examines what motivates researchers to pursue humanoid robotics, and provides a taste of the evolution of this field over time. It summarizes work on legged humanoid locomotion, whole-body activities, and approaches to human–robot communication. It concludes with a brief discussion of factors that may influence the future of humanoid robots.

Footstep planning modeled as a whole-body, inverse-kinematic problem (experiment)

Author  Eiichi Yoshida

Video ID : 600

The whole-body, inverse-kinematic motion including locomotion in video 596 has been experimentally validated by using HPR-2 humanoid robot. The challenging motion-planning problem of picking up an object almost between its feet has been successfully solved with the proposed framework.