The Pransky interview: Dr James Kuffner, CEO at Toyota Research Institute Advanced Development, Coinventor of the rapidly, exploring random tree algorithm

2019 ◽  
Vol 47 (1) ◽  
pp. 7-11
Author(s):  
Joanne Pransky
Keyword(s):  
Motion Planning ◽  
Speech Synthesis ◽  
Industrial Robot ◽  
Humanoid Robots ◽  
Random Tree ◽  
Content Type ◽  
Cloud Robotics ◽  
Engineering Team ◽  
Advanced Development ◽  
Research Institute

Purpose The following article is a “Q&A interview” conducted by Joanne Pransky of Industrial Robot Journal as a method to impart the combined technological, business and personal experience of a prominent, robotic industry PhD-turned entrepreneur regarding his pioneering efforts of bringing technological inventions to market. The paper aims to discuss these issues. Design/methodology/approach The interviewee is Dr James Kuffner, CEO at Toyota Research Institute Advanced Development (TRI-AD). Kuffner is a proven entrepreneur and inventor in robot and motion planning and cloud robotics. In this interview, Kuffner shares his personal and professional journey from conceptualization to commercial realization. Findings Dr Kuffner received BS, MS and PhD degrees from the Stanford University’s Department of Computer Science Robotics Laboratory. He was a Japan Society for the Promotion of Science (JSPS) Postdoctoral Research Fellow at the University of Tokyo where he worked on software and planning algorithms for humanoid robots. He joined the faculty at Carnegie Mellon University’s Robotics Institute in 2002 where he served until March 2018. Kuffner was a Research Scientist and Engineering Director at Google from 2009 to 2016. In January 2016, he joined TRI where he was appointed the Chief Technology Officer and Area Lead, Cloud Intelligence and is presently an Executive Advisor. He has been CEO of TRI-AD since April of 2018. Originality/value Dr Kuffner is perhaps best known as the co-inventor of the rapidly exploring random tree (RRT) algorithm, which has become a key standard benchmark for robot motion planning. He is also known for introducing the term “Cloud Robotics” in 2010 to describe how network-connected robots could take advantage of distributed computation and data stored in the cloud. Kuffner was part of the initial engineering team that built Google’s self-driving car. He was appointed Head of Google’s Robotics Division in 2014, which he co-founded with Andy Rubin to help realize the original Cloud Robotics concept. Kuffner also co-founded Motion Factory, where he was the Senior Software Engineer and a member of the engineering team to develop C++ based authoring tools for high-level graphic animation and interactive multimedia content. Motion Factory was acquired by SoftImage in 2000. In May 2007, Kuffner founded, and became the Director of Robot Autonomy where he coordinated research and software consulting for industrial and consumer robotics applications. In 2008, he assisted in the iOS development of Jibbigo, the first on-phone, real-time speech recognition, translation and speech synthesis application for the iPhone. Jibbigo was acquired by Facebook in 2013. Kuffner is one of the most highly cited authors in the field of robotics and motion planning, with over 15,000 citations. He has published over 125 technical papers and was issued more than 50 patents related to robotics and computer vision technology.

Human skill integrated motion planning of assembly manipulation for 6R industrial robot

2019 ◽  
Vol 46 (1) ◽  
pp. 171-180
Author(s):  
Yi Liu ◽  
Ming Cong ◽  
Hang Dong ◽  
Dong Liu
Keyword(s):  
Deep Learning ◽  
Motion Planning ◽  
Industrial Robot ◽  
Robotic Assembly ◽  
3D Vision ◽  
Content Type ◽  
Human Skill ◽  
Depth Images ◽  
Positioning Method ◽  
Assembly Technology

Purpose The purpose of this paper is to propose a new method based on three-dimensional (3D) vision technologies and human skill integrated deep learning to solve assembly positioning task such as peg-in-hole. Design/methodology/approach Hybrid camera configuration was used to provide the global and local views. Eye-in-hand mode guided the peg to be in contact with the hole plate using 3D vision in global view. When the peg was in contact with the workpiece surface, eye-to-hand mode provided the local view to accomplish peg-hole positioning based on trained CNN. Findings The results of assembly positioning experiments proved that the proposed method successfully distinguished the target hole from the other same size holes according to the CNN. The robot planned the motion according to the depth images and human skill guide line. The final positioning precision was good enough for the robot to carry out force controlled assembly. Practical implications The developed framework can have an important impact on robotic assembly positioning process, which combine with the existing force-guidance assembly technology as to build a whole set of autonomous assembly technology. Originality/value This paper proposed a new approach to the robotic assembly positioning based on 3D visual technologies and human skill integrated deep learning. Dual cameras swapping mode was used to provide visual feedback for the entire assembly motion planning process. The proposed workpiece positioning method provided an effective disturbance rejection, autonomous motion planning and increased overall performance with depth images feedback. The proposed peg-hole positioning method with human skill integrated provided the capability of target perceptual aliasing avoiding and successive motion decision for the robotic assembly manipulation.

The Pransky interview: Dr Jun Ho Oh, Professor and Director of Humanoid Robot Research Center, KAIST; Cofounder, Rainbow Robotics Co.

2017 ◽  
Vol 44 (6) ◽  
pp. 695-699
Author(s):  
Joanne Pransky
Keyword(s):  
Mechanical Engineering ◽  
Humanoid Robot ◽  
Industrial Robot ◽  
Science And Technology ◽  
Vice President ◽  
Humanoid Robots ◽  
Real Time Control ◽  
Commercial Development ◽  
Content Type ◽  
The University

Purpose The following paper is a “Q&A interview” conducted by Joanne Pransky of Industrial Robot Journal as a method to impart the combined technological, business and personal experience of a prominent, robotic industry PhD-turned-entrepreneur regarding the commercialization and challenges of bringing a technological invention to market. This paper aims to discuss these issues. Design/methodology/approach The interviewee is Dr Jun Ho Oh, Professor of Mechanical Engineering at the Korea Advanced Institute of Science and Technology (KAIST) and Director of KAIST’s Hubolab. Determined to build a humanoid robot in the early 2000s to compete with Japan’s humanoids, Dr Oh and KAIST created the KHR1. This research led to seven more advanced versions of a biped humanoid robot and the founding of the Robot for Artificial Intelligence and Boundless Walking (Rainbow) Co., a professional technological mechatronics company. In this interview, Dr Oh shares the history and success of Korea’s humanoid robot research. Findings Dr Oh received his BSc in 1977 and MSc in Mechanical Engineering in 1979 from Yonsei University. Oh worked as a Researcher for the Korea Atomic Energy Research Institute before receiving his PhD from the University of California (UC) Berkeley in mechanical engineering in 1985. After his PhD, Oh remained at UC Berkeley to do Postdoctoral research. Since 1985, Oh has been a Professor of Mechanical Engineering at KAIST. He was a Visiting Professor from 1996 to 1997 at the University of Texas Austin. Oh served as the Vice President of KAIST from 2013-2014. In addition to teaching, Oh applied his expertise in robotics, mechatronics, automatic and real-time control to the commercial development of a series of humanoid robots. Originality/value Highly self-motivated and always determined, Dr Oh’s initial dream of building the first Korean humanoid bipedal robot has led him to become one of the world leaders of humanoid robots. He has contributed widely to the field over the nearly past two decades with the development of five versions of the HUBO robot. Oh led Team KAIST to win the 2015 DARPA Robotics Challenge (DRC) and a grand prize of US$2m with its humanoid robot DRC-HUBO+, beating 23 teams from six countries. Oh serves as a robotics policy consultant for the Korean Ministry of Commerce Industry and Energy. He was awarded the 2016 Changjo Medal for Science and Technology, the 2016 Ho-Am Prize for engineering, and the 2010 KAIST Distinguished Professor award. He is a member of the Korea Academy of Science and Technology.

Peg-In-Hole search using convex optimization techniques

2017 ◽  
Vol 44 (5) ◽  
pp. 618-628 ◽  
Author(s):  
Kamal Sharma ◽  
Varsha Shirwalkar ◽  
Prabir K. Pal
Keyword(s):  
Motion Planning ◽  
Depth Profile ◽  
Industrial Robot ◽  
A Priori ◽  
Optimization Techniques ◽  
Working Environment ◽  
Search Problem ◽  
Content Type ◽  
Priori Information ◽  
Practical Implications

Purpose This paper aims to provide a solution to the first phase of a force-controlled circular Peg-In-Hole assembly using an industrial robot. The paper suggests motion planning of the robot’s end-effector so as to perform Peg-In-Hole search with minimum a priori information of the working environment. Design/methodology/approach The paper models Peg-In-Hole search problem as a problem of finding the minima in depth profile for a particular assembly. Thereafter, various optimization techniques are used to guide the robot to locate minima and complete the hole search. This approach is inspired by a human’s approach of searching a hole by moving peg in various directions so as to search a point of maximum insertion which is same as the minima in depth profile. Findings The usage of optimization techniques for hole search allows the robot to work with minimum a priori information of the working environment. Also, the iterative nature of the techniques adapts to any disturbance during assembly. Practical implications The techniques discussed here are quite useful if a force-controlled assembly needs to be performed in a highly unknown environment and also when the assembly setup can get disturbed in between. Originality/value The concept is original and provides a non-conventional use of optimization techniques, not for optimization of some process directly but for an industrial robot’s motion planning.

Task motion planning for anthropomorphic arms based on human arm movement primitives

2020 ◽  
Vol 47 (5) ◽  
pp. 669-681
Author(s):  
Shiqiu Gong ◽  
Jing Zhao ◽  
Ziqiang Zhang ◽  
Biyun Xie
Keyword(s):  
Motion Planning ◽  
Arm Movement ◽  
Humanoid Robots ◽  
Planning Method ◽  
Service Robots ◽  
Joint Angles ◽  
Content Type ◽  
Structure And Motion ◽  
Human Arm ◽  
Arm Motion

Purpose This paper aims to introduce the human arm movement primitive (HAMP) to express and plan the motions of anthropomorphic arms. The task planning method is established for the minimum task cost and a novel human-like motion planning method based on the HAMPs is proposed to help humans better understand and plan the motions of anthropomorphic arms. Design/methodology/approach The HAMPs are extracted based on the structure and motion expression of the human arm. A method to slice the complex tasks into simple subtasks and sort subtasks is proposed. Then, a novel human-like motion planning method is built through the selection, sequencing and quantification of HAMPs. Finally, the HAMPs are mapped to the traditional joint angles of a robot by an analytical inverse kinematics method to control the anthropomorphic arms. Findings For the exploration of the motion laws of the human arm, the human arm motion capture experiments on 12 subjects are performed. The results show that the motion laws of human arm are reflected in the selection, sequencing and quantification of HAMPs. These motion laws can facilitate the human-like motion planning of anthropomorphic arms. Originality/value This study presents the HAMPs and a method for selecting, sequencing and quantifying them in human-like style, which leads to a new motion planning method for the anthropomorphic arms. A similar methodology is suitable for robots with anthropomorphic arms such as service robots, upper extremity exoskeleton robots and humanoid robots.

Motion planning for humanoid robot dynamically stepping over consecutive large obstacles

2016 ◽  
Vol 43 (2) ◽  
pp. 204-220 ◽  
Author(s):  
Fayong Guo ◽  
Tao Mei ◽  
Minzhou Luo ◽  
Marco Ceccarelli ◽  
Ziyi Zhao ◽  
...  
Keyword(s):  
Motion Planning ◽  
Trajectory Planning ◽  
Humanoid Robot ◽  
Extreme Case ◽  
Humanoid Robots ◽  
Decision Criterion ◽  
Body Motion ◽  
Feasibility Analysis ◽  
Upper Body ◽  
Content Type

Purpose – Humanoid robots should have the ability of walking in complex environment and overcoming large obstacles in rescue mission. Previous research mainly discusses the problem of humanoid robots stepping over or on/off one obstacle statically or dynamically. As an extreme case, this paper aims to demonstrate how the robots can step over two large obstacles continuously. Design/methodology/approach – The robot model uses linear inverted pendulum (LIP) model. The motion planning procedure includes feasibility analysis with constraints, footprints planning, legs trajectory planning with collision-free constraint, foot trajectory adapter and upper body motion planning. Findings – The motion planning with the motion constraints is a key problem, which can be considered as global optimization issue with collision-free constraint, kinematic limits and balance constraint. With the given obstacles, the robot first needs to determine whether it can achieve stepping over, if feasible, and then the robot gets the motion trajectory for the legs, waist and upper body using consecutive obstacles stepping over planning algorithm which is presented in this paper. Originality/value – The consecutive stepping over problem is proposed in this paper. First, the paper defines two consecutive stepping over conditions, sparse stepping over (SSO) and tight stepping over (TSO). Then, a novel feasibility analysis method with condition (SSO/TSO) decision criterion is proposed for consecutive obstacles stepping over. The feasibility analysis method’s output is walking parameters with obstacles’ information. Furthermore, a modified legs trajectory planning method with center of mass trajectory compensation using upper body motion is proposed. Finally, simulations and experiments for SSO and TSO are carried out by using the XT-I humanoid robot platform with the aim to verify the validity and feasibility of the novel methods proposed in this paper.

Motion planning for a new-model obstacle-crossing mobile welding robot

2014 ◽  
Vol 41 (1) ◽  
pp. 87-97 ◽  
Author(s):  
Tao Zhang ◽  
Minghui Wu ◽  
Yanzheng Zhao ◽  
Shanben Chen
Keyword(s):  
Motion Planning ◽  
Welding Speed ◽  
Industrial Robot ◽  
Kinematic Model ◽  
Special Method ◽  
Welding Robot ◽  
Obstacle Crossing ◽  
Content Type ◽  
New Model ◽  
Mobile Welding Robot

Purpose – The purpose of this paper is to discuss motion planning about crossing obstacles and welding trajectory for a new-model mobile obstacle-crossing welding robot system. The robot can cross the obstacle in this way that one of the three adhesion mobile parts is pulled off the ground in turn. An optimal obstacle-crossing approach needs to be studied to improve the welding efficiency. Design/methodology/approach – According to the characteristics of this mobile welding robot, two methods for crossing obstacles are compared. A special method is used for obstacle-crossing and welding. The kinematic model is established. By the optimization method, the optimum parameters for crossing obstacles are calculated. The welding speed when the robot is crossing the obstacle is very important, so its value must be in a certain range. Finally, the tracks of the wheels when the robot is crossing the obstacle are analyzed in order to observe the obstacle-crossing process. Findings – According to the analysis, the maximum speed of the vehicle in the obstacle-crossing is determined. When crossing the obstacle, the robot can do welding simultaneously. The welding speed cannot exceed a certain value. In the obstacle-crossing process, the tracks of the wheels can reflect the process. According to the obtained conclusion, the obstacle-crossing experiments are successfully completed, and the welding effect is good. The results can prove that the proposed method is feasible. Research limitations/implications – The speed of obstacle-crossing is not very large. It has some relationships with the lifting speed of the wheels, which is determined by the quality of drive motor. More efficient robot must be developed to meet the needs of industrial robot. Practical implications – Based on the excellent obstacle-crossing and welding capabilities, the robot with the new mechanism has a widely applying prospect in the field of welding and inspecting large equipment. Originality/value – The obstacle-crossing approach has certain innovation. The way that the robot can maintain continuous welding when crossing the obstacle is of a great significance.

The Pransky interview: Dr Aaron Edsinger, CEO and cofounder at Hello Robot Inc

2018 ◽  
Vol 45 (6) ◽  
pp. 710-714
Author(s):  
Joanne Pransky
Keyword(s):  
Artificial Intelligence ◽  
Computer Science ◽  
San Francisco ◽  
Systems Engineering ◽  
Industrial Robot ◽  
Humanoid Robots ◽  
Massachusetts Institute Of Technology ◽  
Content Type ◽  
Institute Of Technology ◽  
Artificial Intelligence Laboratory

Purpose The purpose of this paper is to present a “Q&A interview” conducted by Joanne Pransky of Industrial Robot Journal as a method to impart the combined technological, business and personal experience of a prominent, robotic industry PhD-turned entrepreneur regarding the evolution, commercialization and challenges of bringing a technological invention to market. Design/methodology/approach The interviewee is Dr Aaron Edsinger, a proven entrepreneur and inventor in the field of human-collaborative robotics. Dr Edsinger shares his journey that led him from developing humanoids at Rodney Brooks’ Computer Science and Artificial Intelligence Laboratory at MIT, to cofounding four companies, two of which got purchased by Google. Findings Dr Edsinger received a BS degree in Computer Systems Engineering from Stanford, an MS in Computer Science from the Massachusetts Institute of Technology (MIT) and a PhD in Computer Science from MIT and did post-doctorate research in the Humanoid Robotics Group at the MIT Computer Science and Artificial Intelligence Lab. He co-founded his first company Meka Robotics in 2007 and that same year, he started his second company, HStar Technologies. In 2011, he cofounded Redwood Robotics, and in 2013, he sold Meka and Redwood to Google. From 2013 to 2017, he was a Robotics Director at Google. In August of 2017, he cofounded Hello Robot Inc. Originality/value Dr Edsinger’s work in robotics grew out of the San Francisco robotic art scene in the 1990s. Since then, he has collaborated and built over a dozen research and artistic robot platforms and has been granted 28 patents. His world-class robotic systems encompass Dr Edsinger’s innovative research in dexterous manipulation in unstructured environments, force controlled compliant actuation, human safe robotics, integrated mechatronic engineering and the design of humanoid robots. Domo, the humanoid robot he built, was named one of Time magazine’s Best Inventions of the Year for 2007. Out of the eight robot companies Google purchased in 2013, two were cofounded by Dr Edsinger. In 2017, Dr Edsinger left Google to cofound his new company, Hello Robot Inc, a stealth mode consumer robot company.

The Pransky interview: Dr Mark W. Tilden, Robotics Physicist

2014 ◽  
Vol 41 (2) ◽  
pp. 113-118
Author(s):  
Joanne Pransky
Keyword(s):  
Personal Experience ◽  
Industrial Robot ◽  
National Laboratory ◽  
Humanoid Robots ◽  
Turn Of The Millennium ◽  
Content Type ◽  
Los Alamos National Laboratory ◽  
Experimental Physicist ◽  
Solar Powered ◽  
The University

Purpose – The following article is a “Q&A interview” conducted by Joanne Pransky of Industrial Robot Journal as a method to impart the combined technological, business, and personal experience of a prominent, robotic industry engineer-turned entrepreneur regarding the evolution, commercialization, and challenges of bringing a technological invention to market. The paper aims to discuss these issues. Design/methodology/approach – The interviewee is Dr Mark W. Tilden, a Robotics Physicist and the inventor of BEAM Robotics. Having built his first 100 bots by the age of nine, Tilden goes on to study at the University of Waterloo and later works at the Los Alamos National Laboratory. There he develops a variety of biomorphic robots including interplanetary explorers and solar-powered bots. During this time, Tilden founds the first BEAM International Olympics. Solarbotics is also formed to disseminate BEAM technologies. At the turn of the millennium, after being approached by toy manufacturer WowWee, Tilden applies his BEAM technology to the consumer toy industry. Findings – From Robobiologist to Chaos Engineer to Toy Consultant to Robotics Physicist, Tilden describes the several decade evolution of his Biomech technologies. Originality/value – The Father of BEAM Robotics, who initially designs single, minimalist biomorphic robots for the space and military industries, transforms his research into the first commercially available affordable humanoid companion for the personal and entertainment robotics industries, culminating in a total of nearly 25 million robots sold worldwide. This experimental physicist continues his pioneering Biomech efforts with hybridization collaborations on life-sized humanoid robots for the home and office.

Effectiveness of multi-gated sequence model for the learning of kinematics and dynamics of an industrial robot

2020 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Aditya Singh ◽  
Padmakar Pandey ◽  
G.C. Nandi
Keyword(s):  
Deep Learning ◽  
Inverse Kinematics ◽  
Inverse Dynamics ◽  
Industrial Robot ◽  
Closed Form Solution ◽  
Humanoid Robots ◽  
Industrial Robots ◽  
Joint Torque ◽  
Data Set ◽  
Content Type

Purpose For efficient trajectory control of industrial robots, a cumbersome computation for inverse kinematics and inverse dynamics is needed, which is usually developed using spatial transformation using Denavit–Hartenberg principle and Lagrangian or Newton–Euler methods, respectively. The model is highly non-linear and needs to deal with uncertainties because of lack of accurate measurement of mechanical parameters, noise and non-inclusion of joint friction, which results in some inaccuracies in predicting accurate torque trajectories. To get a guaranteed closed form solution, the robot designers normally follow Pieper’s recommendation and compromise with the mechanical design. While this may be acceptable for the industrial robots where the aesthetic look is not that important, it is not for humanoid and social robots. To help solve this problem, this study aims to propose an alternative machine learning-based computational approach based on a multi-gated sequence model for finding appropriate mapping between Cartesian space to joint space and motion space to joint torque space. Design/methodology/approach First, the authors generate sufficient data required for the sequence model, using forward kinematics and forward dynamics by running N number of nested loops, where N is the number of joints of the robot. Subsequently, to develop a learning-based model based on sequence analysis, the authors propose to use long short-term memory (LSTM) and hence, train an LSTM model, the architecture details of which have been discussed in the paper. To make LSTM learning algorithms perform efficiently, the authors need to detect and eliminate redundant features from the data set, which the authors propose to do using an elegant statistical tool called Pearson coefficient. Findings To validate the proposed model, the authors have performed rigorous experiments using both hardware and simulation robots (Baxter/Anukul robot) available in their laboratory and KUKA simulation robot data set made available from Neural Learning for Robotics Laboratory. Through several characteristic plots, it has been shown that a sequence-based LSTM model of deep learning architecture with non-redundant features could help the robots to learn smooth and accurate trajectories more quickly compared to data sets having redundancy. Such data-driven modeling techniques can change the future course of direction of robotics research for solving the classical problems such as trajectory planning and motion planning for manipulating industrial as well as social humanoid robots. Originality/value The present investigation involves development of deep learning-based computation model, statistical analyses to eliminate redundant features, data creation from one hardware robot (Anukul) and one simulation robot model (KUKA), rigorously training and testing separately two computational models (specially configured two LSTM models) – one for learning inverse kinematics and one for learning inverse dynamics problem – and comparison of the inverse dynamics model with the state-of-the-art model. Hence, the authors strongly believe that the present paper is compact and complete to get published in a reputed journal so that dissemination of new ideas can benefit the researchers in the area of robotics.

The parameter identification model considering both geometric parameters and joint stiffness

2019 ◽  
Vol 47 (1) ◽  
pp. 76-81
Author(s):  
Jing Bai ◽  
Le Fan ◽  
Shuyang Zhang ◽  
Zengcui Wang ◽  
Xiansheng Qin
Keyword(s):  
Parameter Identification ◽  
Industrial Robot ◽  
Joint Stiffness ◽  
Geometric Parameters ◽  
Least Square ◽  
Experimental Result ◽  
Positional Accuracy ◽  
Content Type ◽  
The Absolute ◽  
Identification Model

Purpose Both geometric and non-geometric parameters have noticeable influence on the absolute positional accuracy of 6-dof articulated industrial robot. This paper aims to enhance it and improve the applicability in the field of flexible assembling processing and parts fabrication by developing a more practical parameter identification model. Design/methodology/approach The model is developed by considering both geometric parameters and joint stiffness; geometric parameters contain 27 parameters and the parallelism problem between axes 2 and 3 is involved by introducing a new parameter. The joint stiffness, as the non-geometric parameter considered in this paper, is considered by regarding the industrial robot as a rigid linkage and flexible joint model and adds six parameters. The model is formulated as the form of error via linearization. Findings The performance of the proposed model is validated by an experiment which is developed on KUKA KR500-3 robot. An experiment is implemented by measuring 20 positions in the work space of this robot, obtaining least-square solution of measured positions by the software MATLAB and comparing the result with the solution without considering joint stiffness. It illustrates that the identification model considering both joint stiffness and geometric parameters can modify the theoretical position of robots more accurately, where the error is within 0.5 mm in this case, and the volatility is also reduced. Originality/value A new parameter identification model is proposed and verified. According to the experimental result, the absolute positional accuracy can be remarkably enhanced and the stability of the results can be improved, which provide more accurate parameter identification for calibration and further application.

Sign in / Sign up

Export Citation Format

Share Document