This data set is captured from a robot workcell that is performing activities representative of several manufacturing operations. The workcell contains two, 6-degree-of-freedom robot manipulators where one robot is performing material handling operations (e.g., transport parts into and out of a specific work space) while the other robot is performing a simulated precision operation (e.g., the robot touching the center of a part with a tool tip that leaves a mark on the part). This precision operation is intended to represent a precise manufacturing operation (e.g., welding, machining). The goal of this data set is to provide robot level and process level measurements of the workcell operating in nominal parameters. There are no known equipment or process degradations in the workcell. The material handling robot will perform pick and place operations, including moving simulated parts from an input area to in-process work fixtures. Once parts are placed in/on the work fixtures, the second robot will interact with the part in a specified precise manner. In this specific instance, the second robot has a pen mounted to its tool flange and is drawing the NIST logo on a surface of the part. When the precision operation is completed, the material handling robot will then move the completed part to an output. This suite of data includes process data and performance data, including timestamps. Timestamps are recorded at predefined state changes and events on the PLC and robot controllers, respectively. Each robot controller and the PLC have their own internal clocks and, due to hardware limitations, the timestamps recorded on each device are relative to their own internal clocks. All timestamp data collected on the PLC is available for real-time calculations and is recorded. The timestamps collected on the robots are only available as recorded data for post-processing and analysis. The timestamps collected on the PLC correspond to 14 part state changes throughout the processing of a part. Timestamps are recorded when PLC-monitored triggers are activated by internal processing (PLC trigger origin) or after the PLC receives an input from a robot controller (robot trigger origin). Records generated from PLC-originated triggers include parts entering the work cell, assignment of robot tasks, and parts leaving the work cell. PLC-originating triggers are activated by either internal algorithms or sensors which are monitored directly in the PLC Inputs/Outputs (I/O). Records generated from a robot-originated trigger include when a robot begins operating on a part, when the task operation is complete, and when the robot has physically cleared the fixture area and is ready for a new task assignment. Robot-originating triggers are activated by PLC I/O. Process data collected in the workcell are the variable pieces of process information. This includes the input location (single option in the initial configuration presented in this paper), the output location (single option in the initial configuration presented in this paper), the work fixture location, the part number counted from startup, and the part type (task number for drawing robot). Additional information on the context of the workcell operations and the captured data can be found in the attached files, which includes a README.txt, along with several noted publications. Disclaimer: Certain commercial entities, equipment, or materials may be identified or referenced in this data, or its supporting materials, in order to illustrate a point or concept. Such identification or reference is not intended to imply recommendation or endorsement by NIST; nor does it imply that the entities, materials, equipment or data are necessarily the best available for the purpose. The user assumes any and all risk arising from use of this dataset.

This data set is captured from a robot workcell that is performing activities representative of several manufacturing operations. The workcell contains two, 6-degree-of-freedom robot manipulators where one robot is performing material handling operations (e.g., transport parts into and out of a specific work space) while the other robot is performing a simulated precision operation (e.g., the robot touching the center of a part with a tool tip that leaves a mark on the part). This precision operation is intended to represent a precise manufacturing operation (e.g., welding, machining). The goal of this data set is to provide robot level and process level measurements of the workcell operating in nominal parameters. There are no known equipment or process degradations in the workcell. The material handling robot will perform pick and place operations, including moving simulated parts from an input area to in-process work fixtures. Once parts are placed in/on the work fixtures, the second robot will interact with the part in a specified precise manner. In this specific instance, the second robot has a pen mounted to its tool flange and is drawing the NIST logo on a surface of the part. When the precision operation is completed, the material handling robot will then move the completed part to an output. This suite of data includes process data and performance data, including timestamps. Timestamps are recorded at predefined state changes and events on the PLC and robot controllers, respectively. Each robot controller and the PLC have their own internal clocks and, due to hardware limitations, the timestamps recorded on each device are relative to their own internal clocks. All timestamp data collected on the PLC is available for real-time calculations and is recorded. The timestamps collected on the robots are only available as recorded data for post-processing and analysis. The timestamps collected on the PLC correspond to 14 part state changes throughout the processing of a part. Timestamps are recorded when PLC-monitored triggers are activated by internal processing (PLC trigger origin) or after the PLC receives an input from a robot controller (robot trigger origin). Records generated from PLC-originated triggers include parts entering the work cell, assignment of robot tasks, and parts leaving the work cell. PLC-originating triggers are activated by either internal algorithms or sensors which are monitored directly in the PLC Inputs/Outputs (I/O). Records generated from a robot-originated trigger include when a robot begins operating on a part, when the task operation is complete, and when the robot has physically cleared the fixture area and is ready for a new task assignment. Robot-originating triggers are activated by PLC I/O. Process data collected in the workcell are the variable pieces of process information. This includes the input location (single option in the initial configuration presented in this paper), the output location (single option in the initial configuration presented in this paper), the work fixture location, the part number counted from startup, and the part type (task number for drawing robot). Additional information on the context of the workcell operations and the captured data can be found in the attached files, which includes a README.txt, along with several noted publications. Disclaimer: Certain commercial entities, equipment, or materials may be identified or referenced in this data, or its supporting materials, in order to illustrate a point or concept. Such identification or reference is not intended to imply recommendation or endorsement by NIST; nor does it imply that the entities, materials, equipment or data are necessarily the best available for the purpose. The user assumes any and all risk arising from use of this dataset.

NIST is developing metrics and test methods to benchmark the performance of robotic systems when performing manufacturing tasks. The ability to perform simple insertions is critical for robotic systems in manufacturing. A simple peg-in-hole test was designed to measure a robotic system's capability for performing these simple insertions. The dataset captures the performance metrics of a robotic system outfitted with a robotic hand and a robotic gripper to study the effect of next-generation robotic hand technology versus conventional parallel gripper technologies.

고해상도, 저해상도 이미지 데이터 및 로봇 핸드를 활용한 객체 데이터

실제와 가상환경에서 취득한 RGB-D이미지와 그에 대응하는 로봇관절각도, End-effector 포즈, 힘/토크, 촉각데이터로 구성 된 로봇 그리퍼로 200종 이상의 물체를 조작하는 대규모 물리 환경 로봇 조작 데이터

An advanced approach to flexible manufacturing is to move robotic manipulators (also referred to as industrial arms), using an AGV or mobile robot, between workstations. This integrated system is referred to as a mobile manipulator. Prior to industrial acceptance and standards development for mobile manipulators, users of these new systems will expect manufacturers to provide real performance data to guide their procurement and assure suitability for given application tasks. A test method that uses an artifact, called the Reconfigurable Mobile Manipulator Artifact (RMMA), is described in [Bostelman RV, Li-Baboud Y, Legowik S, Hong TH, Foufou S., "Mobile Manipulator Performance Measurement Data". 2017 Jun 27] and compared to an optical tracking system that was used as ground truth for the RMMA and mobile manipulator. Measurement data of an AGV, an onboard robot arm, and an optical tracking system were recorded and are described in the paper and are available for download using the link available in this record. The data needed to make these three measurements was collected during two tests; both tests have corresponding timestamps relative to global positioning system (GPS) time, where the computer clocks are synchronized using the Network Time Protocol. It is expected that the user of the information within the paper and the data files will have sufficient knowledge to implement mobile manipulation testing and evaluation.

An advanced approach to flexible manufacturing is to move robotic manipulators (also referred to as industrial arms), using an AGV or mobile robot, between workstations. This integrated system is referred to as a mobile manipulator. Prior to industrial acceptance and standards development for mobile manipulators, users of these new systems will expect manufacturers to provide real performance data to guide their procurement and assure suitability for given application tasks. A test method that uses an artifact, called the Reconfigurable Mobile Manipulator Artifact (RMMA), is described in [Bostelman RV, Li-Baboud Y, Legowik S, Hong TH, Foufou S., "Mobile Manipulator Performance Measurement Data". 2017 Jun 27] and compared to an optical tracking system that was used as ground truth for the RMMA and mobile manipulator. Measurement data of an AGV, an onboard robot arm, and an optical tracking system were recorded and are described in the paper and are available for download using the link available in this record. The data needed to make these three measurements was collected during two tests; both tests have corresponding timestamps relative to global positioning system (GPS) time, where the computer clocks are synchronized using the Network Time Protocol. It is expected that the user of the information within the paper and the data files will have sufficient knowledge to implement mobile manipulation testing and evaluation.

자율주행 로봇의 실내외 통합 주행을 위한 주행 데이터 구축날씨 (맑음, 흐림, 비)와 아침/낮/저녁 등의 다양한 환경에서 실내 및 실외에서 로봇이 주행할 때 보이는 여러 객체들의 종류들과 객체 위치들의 데이터를 제공하여 실내 및/혹은 실외공간, 통합 공간 기반의 다양한 연구를 할 수 있는 데이터 제공

The ability to utilize prognostic system health information in operational decision making, especially when fused with information about future operational, environmental, and mission requirements, is becoming desirable for both manned and unmanned aerospace vehicles. A vehicle capable of evaluating its own health state and making (or assisting the crew in making) decisions with respect to its system health evolution over time will be able to go further and accomplish more mission objectives than a vehicle fully dependent on human control. This paper describes the development of a hardware testbed for integration and testing of prognostics-enabled decision making technologies. Although the testbed is based on a planetary rover platform (K11), the algorithms being developed on it are expected to be applicable to a variety of aerospace vehicle types, from unmanned aerial vehicles and deep space probes to manned aircraft and spacecraft. A variety of injectable fault modes is being investigated for electrical, mechanical, and power subsystems of the testbed. A software simulator of the K11 has been developed, for both nominal and off-nominal operating modes, which allows prototyping and validation of algorithms prior to their deployment on hardware. The simulator can also aid in the decision-making process. The testbed is designed to have interfaces that allow reasoning software to be integrated and tested quickly, making it possible to evaluate and compare algorithms of various types and from different sources. Currently, algorithms developed (or being developed) at NASA Ames - a diagnostic system, a prognostic system, a decision-making module, a planner, and an executive - are being used to complete the software architecture and validate design of the testbed.