Special report on the robot industry: the evolution trend of industrial robots from ABB

Industrial robots: gradually moving towards intelligence/generalization, and the prospects for industrial application are promising

Development history: industrial robots are moving towards intelligent/generalized development

The development process of industrial robots can be divided into three stages: Technology embryonic stage (50-60s of the 20th century, the first generation of industrial robots): the prototype of industrial robots during the Second World War, which was originally out of the development needs of the military and nuclear industries, and was used to meet the remote-controlled manipulators that replaced human beings for radioactive material processing. Since then, United States inventor George · Devor first proposed the concept of industrial robots and applied for a patent for "editable articulated material transfer device" in 1954. In 1958, the world's first robotics company, Unimation, was founded, and in 1959, using George ·'s patent, the world's first true industrial robot, Unimate, was launched, ushering in a new era of robotics development. During this period, it was mainly the first generation of industrial robots, that is, teach-in reproduction robots that did not have autonomous decision-making ability and environmental perception ability.

Industrial rise stage (late 20th century, second generation of industrial robots): Since 1968, Unimation has successively transferred robot manufacturing technology to Kawasaki in Japan and GKN in United Kingdom, and the industrial robot business has expanded to Japan and Europe, and has begun to move towards industrialization and globalization. In 1969, ASEA developed the world's first spraying robot, and Kawasaki successfully launched Japan's first hydraulically powered robot, the Kawasaki Unimate 2000. Since then, more companies have entered the field of industrial robots, forming a pattern of European robot companies represented by ABB and Kuka and Japan robot enterprises represented by Kawasaki, Fanuc and Yaskawa, and the robot market has grown steadily. During this period, industrial robots have evolved from the first generation to the second generation, capable of perceiving external information and performing simple processing and feedback.

Industrial upgrading stage (early 21st century, the third generation of industrial robots): With the development of industrial automation, industrial robots continue to expand application scenarios and develop core technologies. 1) From the perspective of application scenarios, industrial robots can achieve faster speed, higher accuracy, and a wider range of sizes, models and loads, and realize more extensive and intelligent applications in large-scale workpiece handling production, logistics and transportation, food and beverage, biopharmaceutical, automobile manufacturing, etc. 2) From the perspective of technological development, the core technology of industrial robots has developed rapidly. In 2002 United States Boston Company and Japan Company jointly applied for the first patent for the "Robot Dog" intelligent military robot; In 2004, both Yaskawa and ABB developed controllers that can control multiple robots simultaneously; In 2006, Comau in Italy launched the first wireless teach pendant; In 2015, ABB launched the world's first truly collaborative robot, YuMi. During this period, industrial robots have entered the third generation, with intelligent characteristics, and can use various sensors and measuring instruments to obtain information and use intelligent technology to identify, understand and feedback.

The four families of industrial robots occupy a high position by virtue of their leading advantages in the original business and a high degree of mastery of robot technology. Founded in 1988 as a result of the merger of two well-known electrical companies with a history of 100 years, ASEA and BBC, a high-voltage electrical transmission company, has since become a technology leader in the field of electrification and automation, offering comprehensive and increasingly digital electrification, motion and automation solutions. Since the launch of the world's first microcomputer-controlled robot in 1974, ABB has continuously made breakthrough innovations, continuously improved the product layout of industrial robots in the industrial chain, and improved performance through technological innovation.

Market status: the steady growth market occupied by the "four families".

The scale of the industrial robot market is growing steadily, and the global new installed capacity is growing rapidly. According to IFR data, the global installation of new industrial robots in 2021 was 517,000 units, a year-on-year increase of 31.2%, and a compound growth rate of 12.0% from 2012 to 2021. According to Inkwood Research data, the global industrial robot market will maintain steady growth in the future, from $55.37 billion in 2020 to $165.35 billion in 2028, with a compound growth rate of 14.7% from 2021 to 2028.

The industrial robot market is highly concentrated, dominated by the "four families" of robots. According to IFR data, in the global industrial robot market in 2020, Fanuc, ABB, Yaskawa, and Kuka accounted for more than 65%, of which Fanuc ranked first with 17.3%, ABB second with 15.7%, and Yaskawa and Kuka accounted for 12.9% and 12.1% respectively. The "four families" have different main businesses and core technologies in the field of robotics, with Yaskawa and KUKA focusing on robotics, while Fanuc and ABB start in factory automation equipment and power business, respectively. Among them, ABB's business is centered on electrical, motion control, process automation, robotics and discrete automation, accounting for 47.9% of the revenue of the electrical business and 10.8% of the robot and automation business in 2022.

Driving force: The contraction of population supply and the release of downstream demand drive the growth of the industry

The aging of the population has led to the rise of labor costs, and machines are the general trend for people. The fertility rate of countries around the world has been declining for many years, and according to the National Bureau of Statistics, the proportion of China's population aged 15-64 has dropped to 68.3% in 2021. According to the National Bureau of Statistics, the per capita wage of employees in China's manufacturing industry has reached 86,933 yuan in 2022, a year-on-year increase of 5.2%, higher than the GDP growth rate. The long-term trend of rising labor costs will become a powerful driving force for the development of the industrial robot industry.

The outbreak of new energy vehicles has brought demand release for industrial robots. Driven by new energy vehicles, the automotive industry has entered the next round of expansion cycle, and industrial robots will continue to benefit. Taking China as an example, according to Flush data, the sales of new energy vehicles have entered a stage of explosive growth since 2021, with sales exceeding 800,000 units in June 2023. Semiconductor, photovoltaic, lithium battery and other industries are also expected to bring a large demand for industrial robots. Automotive and 3C electronics are the two industries with the largest demand for industrial robots, and now the demand for 3C electronics is relatively weak, and most companies have begun to actively explore performance growth points outside of automobiles and 3C, such as semiconductors, photovoltaics, lithium batteries, etc. The global semiconductor market will reach $573.5 billion in 2022, and the expansion of semiconductor production is expected to drive the demand for industrial robot construction as AI drives the demand for computing chip design, manufacturing, packaging and other industries.

IT architecture is the core of intelligence, and the importance of software is expected to gradually increase

Industrial robot industry chain: it can be divided into three links: upstream core components, midstream robot body and software, and downstream system integration and application. The upstream core components comprise sensor, controller, reducer, servo system, and the sensor is mainly divided into internal sensor and external sensor, and the servo system mainly comprises two components of servo drive and servo motor; The midstream includes robot ontology manufacturing and supporting software products; The downstream includes system integration and application, system integration is a combination of various robots, sensors, controllers, software, etc. to form a complete automated production line, industrial robots are widely used, of which the main application areas are automobile manufacturing, electronic and electrical, food manufacturing, metal processing and pharmaceutical industry.

The upstream core components are the main cost source of industrial robots, and the profitability is strong. From the cost point of view, in 2020, 60% of the cost of industrial robots will be composed of upstream core components, of which reducers, servo systems, and controllers will account for 35%, 25%, and 10% respectively, and the robot itself will only account for 15%. In 2020, the gross profit margins of reducers, servo systems, and controllers will be about 40%, 35%, and 25%, respectively, and the gross profit margin of midstream ontology manufacturing will be lower than only 15%, and the gross profit margin of downstream system integration will be higher at 35%.

Special report on the robot industry: the evolution trend of industrial robots from ABB

Taking ABB as an example, according to the specific functional structure, we can summarize the core components of industrial robots as: perception layer, execution layer, and decision-making layer. Specifically:

Perception layer: The initial source of information for industrial robots

The sensor is the "sensing organ" of the robot, and it is the source of information for the robot to interact with the environment and achieve precise movement. Sensors are used to sense and collect various information, and convert the measured information into useful signals that are convenient for transmission and processing according to certain rules. The sensor is usually composed of a sensing element and a conversion element, wherein the sensing element refers to the part of the sensor that directly senses the measurement, and the conversion element refers to the part that the sensor can convert the output of the sensing element into an electrical signal suitable for transmission and processing. Sensors are generally divided into internal and external sensors, and the internal sensors mainly measure the internal system status of the robot, such as temperature, speed, voltage, etc.; External sensors mainly measure the information of the external environment, including position sensors, speed sensors, force sensors, vision sensors, etc.

The market size of robot sensors is growing steadily, and the market concentration is low. According to Mordor Intelligence, the global robotics sensor market size is expected to grow from $616 million in 2023 to $924 million by 2028, at a CAGR of 8.4% during 2024-2028. The main players in the robot sensor market are from many companies in different countries around the world, including Switzerland Baumer Group, Tyco Electronics; United States ATI, Tekscan, Futek, Sensata, Honeywell; Japan Fanuc, TDK, Omron Group, etc.

Force sensors are important sensors for industrial robots, and ABB's integrated force control technology is one of the main technologies for force sensors. Most of the industrial robots are in the form of robotic arms, which are composed of the power joints and connecting rods of the robot, and the force sensor can monitor and feedback the force on the robotic arm in real time. With the development of robots to intelligence, the demand for compliant control of interaction with the outside world is higher, and higher requirements are put forward for force control technology. In order to improve the flexibility and accuracy of the sensors, ABB successfully developed integrated force control technology in 2014, based on which ABB offers three integrated force control sensors that are compatible with most of its industrial robots from IRB 140 to IRB 6700. Robots equipped with this technology can respond to subtle changes in the manufacturing process based on external real-time feedback signals, handle fragile items or perform precision tasks like humans, and shorten programming time.

Execution layer: determines the accuracy of the actual movement of the industrial robot

The "muscle" of the robot - the servo system

The servo system is the "muscle" of the robot, an automatic control system that is able to follow or reproduce a process precisely. The servo system consists of a driver, a drive motor, a servo encoder and supporting software. Servo drive controls servo motor, and its function is to convert the control signal into the drive signal of motor movement, generally control servo motor by position, speed and torque three modes, realize the high-precision transmission system positioning. Servo motor is the engine that controls mechanical elements to operate, and after receiving the driving signal of servo drive, converts electrical energy into mechanical energy, that is, converts voltage signal into torque and speed to drive control object. The servo encoder is generally installed at the end of the servo motor, is used to measure the rotation angle and speed, is the signal feedback device of the servo system, and has a joint effect on the accuracy of the servo system.

ABB's servo drives and servo motors offer superior performance and cover a wide range of voltages and powers. ABB's servo drives are designed to meet most voltage ranges and provide high positioning accuracy. The MicroFlex e190 and MotiFlex e180 are highly integrated with Ethernet's flexible connectivity and motor feedback technology and are optimized for demanding motion applications. ABB servo motors also have a complete product range with power coverage for most applications, high speed and high torque for high dynamic response and precise control.

ABB offers specialized selection and commissioning tools for its servo products. Servosize is a configurator for ABB's servo products and supports MotiFlexe180, MicroFlexe190 and E530 servo drives, as well as eSM, DSM and HDS servo motors. MintWorkbench, ServoComposer and AutomationBuilder are programming and debugging tools for ABB products. MintWorkbench can be used for programming and commissioning of MotiFlexe180 and MicroFlexe190 servo drives, ServoComposer can be used for commissioning of E530 servo drives, and AutomationBuilder can be used for programming and commissioning motion control PLCs.

The "joint" of the robot - the reducer

The "joint" of the reducer robot is an intermediate device connecting the servo motor and the actuator. The reducer converts the power of the servo motor at high speed into the mechanical motion of low speed and high torque and transmits it to the actuator through gear meshing, and plays the role of matching speed and transmitting torque. The output speed, transmission accuracy, output torque, etc. of reducer determine the movement speed, positioning accuracy and bearing capacity of the corresponding motion shaft. Generally speaking, each joint of an industrial robot needs to be equipped with a reducer, and most of the industrial robots are six axes, so generally an industrial robot needs six reducers. Reducer can be divided into harmonic reducer and RV reducer, harmonic reducer is simple to use, high precision, but the structural rigidity is not as good as RV reducer, Japan Hamonaco is the world's largest harmonic reducer manufacturer; RV reducer structure is rigid, output torque is large, but its internal structure is more complex, manufacturing cost is high, transmission accuracy is slightly lower, Japan Nabtesco is the world's largest RV reducer manufacturer.

The degree of self-production of reducers in industrial robot enterprises is low, and most of them are outsourced. The reducer market concentration is high, according to the "China Robot Industry Development Report" released by Harbin Institute of Technology in 2021, Japan Harmony Naknaco, Nabtesco and Sumitomo occupy 85% of the share, of which Nabtesco is the inventor of RV reducer, accounting for 60% of the global industrial robot reducer market, and 80% of the RV reducer market of CNC machine tool automatic tool change (ATC) device, the four major families of industrial robots have weak self-production capabilities, all of which purchase Nabtesco reducers. Japan Harmony is the world's first company to develop and produce harmonic reducers, accounting for 15% of the second reducer market.

Decision-making level: the key to determining the performance of industrial robots

The controller plays the role of "cerebellum" in the structure of the robot and is the main factor determining the performance of the robot. The controller receives the sensing information and generates the corresponding control instruction accordingly, controls the robot to complete the action task. Controller comprises two parts of hardware and software, and hardware is control computer and teach pendant, wherein teach pendant is the handheld device that carries out robot manual manipulation and configuration and programming; Software is mainly the application software and system with control algorithm as the core. The working principle of the controller is to compare the input signal of the detection unit with the set value signal, and the deviation signal is calculated, and then the calculation result is output to the actuator, and the actuator changes the operation variable and then outputs to the controlled object.

In the case of ABB, the OmniCore industrial robot controller delivers powerful performance for industrial robots. The C series is the compact controller of the OmniCore family, which is not only greatly reduced in size, but also has flexible integration capabilities; The E-Series is an ultra-compact controller designed to meet the needs of customers running tight electronics assembly lines; The V-Series (Extra Tough) is a medium-sized controller in the modular OmniCore controller family, optimized for medium and large robots in industrial environments, e.g. MH/MT, arc welding.

ABB's IRC5 is the industry standard for robot controller technology. The ABB IRC5 controller is available in single-cabinet, double-cabinet and compact versions, with the single-cabinet IRC5 consisting of two hardware parts: the controller and the FlexPendant teach pendant. The robot can be programmed by the user on the FlexPendant Teach Pendant or on the RobotStudio on the FlexPendant Teach Pendant. In addition, the IRC5 is designed to be flexible and modular. The modular design is more flexible than a centralized controller, and the IRC5 consists of a control module and a drive module, with the option of adding a process module, allowing the IRC5 to flexibly control a six-axis robot and servo-driven workpiece positioners. The control module has its own host computer, which is capable of executing advanced control algorithms, performing composite path calculations for up to 36 servo axes, and directing four drive modules. In addition, IRC5 offers flexibility, security, modularity, application programming interfaces, multi-robot control, and PC tool support.

ABB controllers are supported by their superior motion control technology. QuickMove, TrueMove and MultiMove are ABB's core motion control technologies, which play an important role in improving the robot's performance metrics such as accuracy, speed, cycle time, programmability and synchronization with external devices. TrueMove guarantees the robot's path accuracy, ensuring that the robot's motion path follows the programmed path regardless of speed or operating mode, even after safety protection stops, process stops, program stops, or power is lost; QuickMove can realize motion control in a short time, and can automatically optimize the acceleration and speed of the axis according to different work tasks, so as to achieve more efficient and accurate motion control, improve the production efficiency and accuracy of the robot, and the cycle speed is 25% faster than that of competitors; The introduction of MultiMove further strengthens ABB's leadership in advanced robotic systems and capabilities, enabling up to four robots to work in tandem.

The B&R control system provides a powerful integrated architecture for ABB robots. According to the company's website, in April 2017, ABB announced the acquisition of B&R, a leader in the industrial software industry, a true leader in PLC, industrial PC, and servo control, with a complete integrated architecture. In 2019, ABB and B&R jointly developed a solution to integrate ABB robots into B&R's machine control system. This scheme simplifies the programming of the robot, realizes the parallel work of the workbench and the robot, improves the synchronization of the robot motion and the sensor, realizes the simulation of the robot, and makes the development more convenient.

IT architecture: Software is expected to rise in a step-by-step manner

In the case of ABB, the IT architecture of industrial robots is the base of the decision-making system. In the decision-making system of industrial robots, we see that the controller plays an important role in scheduling robot behavior and executing complex instructions. From the perspective of dismantling the decision-making system, the underlying basic control hardware and control software are the core foundation in the IT architecture of industrial robots.

Hardware: The main control device consists of core hardware such as the DSQC1000 main computer. The single-cabinet IRC5 is equipped with a DSQC1000 main computer, which is the brain of the entire controller, which is used to receive robot motion data and peripheral signals, and send signals to each unit; Axis computer receives the position data that robot axis serial measuring board returns, and compares and calculates with the original position data and outputs new position and speed control parameters; The main servo drive is used to drive the servo motors distributed on the 6 axes of the robot; The I/O module is used to connect external signals and is a user-defined input and output signal board. The safety panel is used to connect safety-relevant signals, and the LED lights remain lit during normal operation.

Software: It generally consists of an operating system and an algorithm library. The operating systems mainly include VxWor (Embedded Real-Time Operating System) and the widely used open source operating systems Windows CE, Embedded Linux, Embedded Linux, μC/OS-II, and the general ROS platform. Wherein algorithm library comprises underlying algorithm library and application process algorithm library, and underlying algorithm library is the algorithm that is developed directly on computer operating system, and control algorithm is used to plan the motion point position of robot, and dynamic algorithm is used to identify the moment of inertia of fuselage load; The application process algorithm library is an algorithm developed by software users. RobotWare is a dedicated software family for ABB's embedded industrial robot controllers that controls a wide range of industrial robots manufactured by ABB. The optional plug-ins provide the robot user with a rich range of system functions, such as multi-task parallelism, transmission of file information to the robot, communication with external systems, advanced motion tasks, etc.

ABB also offers a variety of software packages with different functions for different application scenarios. Including spraying packages, loading and unloading packages, picking packages, machining packages, cutting packages, palletizing packages, PickMaster picking and packing packages, etc. It provides customized software services for welding, cutting, picking, palletizing and other special links to improve the work efficiency of specific production links.

In addition, the programming methods of industrial robots generally include teach-in programming and virtual simulation programming. Teaching programming needs to be guided by the robot step by step, the action that needs to be completed is informed by the teaching pendant operation, the program is generated just after the teaching operation is completed, and the robot can reproduce all the teaching actions according to the generated program, and is suitable for repeated operation robots such as automatic production lines. Virtual simulation programming is to realize programming and simulation two functions through programming software, not only can write programs, but also can carry out simulation to robot motion to verify the effectiveness of the program. Virtual simulation programming requires robot-specific programming software, such as ABB's Robot Studio. Robot Studio is ABB's robot simulation and programming software, which provides a visual programming environment to make robot programming more flexible and efficient, and on the other hand, powerful simulation functions can help users develop and test robot applications. In January 2023, ABB enhanced its RobotStudio robot programming and simulation software with cloud capabilities, the new RobotStudio Cloud enables individuals and teams to remotely collaborate on robot cell design in real-time on any device.

IT architecture: Software is expected to rise in a step-by-step manner. We believe that with the development of industrial robots to intelligence, they have gradually entered the software-defined stage, and robots are no longer a mechanical manufacturing product (slow iteration/high coupling of software and hardware), but gradually evolved into IT products (iterative fast/software and hardware decoupling). Whether it is an operating system, an algorithm library, or programming development software, it is playing an increasingly important role in the process of robot iteration, and the difference in software algorithms will have an important impact on the level of intelligence that the robot can finally achieve. Based on this, we are optimistic about the development opportunities of IT software under the transformation of the robot industry, and the status of software is expected to gradually improve.

AI empowerment is expected to help accelerate industrial transformation

Industry trends: Industrial robots are evolving to embodied intelligence

ChatGPT transforms humans from in-the-loop to on-the-loop in bot flows. Currently, bot pipelines require a dedicated engineer in-the-loop to write code to improve the process. The introduction of ChatGPT can replace the position of humans in the loop, and human (technical or non-technical) users can interact with the language model in the form of on-the-loop through high-level language commands, so as to achieve seamless deployment of various platforms and tasks.

Human users evaluate the quality and safety of ChatGPT's output in a bot pipeline. The main tasks of humans in the bot pipeline are: 1) defining high-level bot function libraries. On the one hand, the high-level robot function library is oriented to the robot platform, which can invoke and guide robot-related actions; On the other hand, for ChatGPT, ensure that the naming of the functions in the library should be easy for ChatGPT to understand and follow. 2) Build a prompt. Pompt describes the task objectives and identifies the functions in the high-level library that ChatGPT is allowed to use. In addition, it can also include constraint information, or tell ChatGPT how to organize its responses. 3) Analyze and evaluate ChatGPT output results and give feedback. In the form of on-the-loop, users evaluate the code output by ChatGPT through direct analysis or simulation, and provide feedback to ChatGPT on the quality and safety of the output code. 4) Iterative. Iterate on the results generated by ChatGPT until they meet human expectations and ensure that the final code can be deployed to the bot for execution.

Simple tasks: ChatGPT is capable of solving simple bot tasks in a zero-shot manner. For simple robot tasks, users only need to provide prompts and function library descriptions, and do not need to provide specific code examples, ChatGPT can zero-shot solve problems such as spatiotemporal reasoning, control of real drones, and industrial detection of drones. 1) Spatiotemporal reasoning: ChatGPT is asked to control a planar robot and capture the basketball position with a visual servo. 2) Real-world drone flight: Use ChatGPT and API to control a real drone and complete object finding tasks. 3) AirSim Industrial Inspection: Based on the AirSim simulator, ChatGPT is used to control the drone in the simulation domain for industrial inspection.

Complex tasks: With on-the-loop human interaction, ChatGPT is able to complete more complex bot control tasks. For more complex problems, ChatGPT cannot be zero-shot or has limited completion effect, at this time, human users can assist ChatGPT in the form of text feedback interaction to complete tasks such as course learning and AirSim obstacle avoidance. 1) Lessons: Teach ChatGPT simple pick-and-place object skills, and use the skills learned in logical combinations for more complex block arrangement tasks. 2) AirSim obstacle avoidance: ChatGPT has built most of the key modules of the obstacle avoidance algorithm, but it still needs some information such as the direction of the drone to be fed back by humans. Human feedback is all high-level natural language, but ChatGPT is able to understand and make code corrections where appropriate.

Implementation: ABB and other leading manufacturers are actively embracing changes in the industry

AI applications in the field of industrial robots are still in the exploratory stage, enabling some intelligent identification and intelligent programming. Take ABB as an example, empowering the e-commerce logistics sector with its AI-powered robotic item picker. According to Statista, global e-commerce revenue is expected to grow by more than 50% over the next five years, from €1.7 trillion in 2019 to €2.6 trillion in 2024. In February 2020, ABB and Silicon Valley AI startup Covariant announced a partnership to bring AI robotic solutions to market, according to the company's website. Covariant's covariant brain is a general artificial intelligence that allows robots to observe, reason, and act in the world around them, completing tasks that are too complex and diverse for traditional programming robots. Covariant's software enables robots to perform reinforcement learning: adapting themselves to new tasks through trial and error, thus constantly expanding the range of objects they can pick and choose. ABB and Covariant AI support solutions were first deployed at Active Ants, a leading provider of e-commerce services in the Netherlands.

AI and machine vision enhancements give robots human dexterity and improve the efficiency and speed of intelligent picking recognition. According to the company's official website, on April 25, 2023, ABB has launched artificial intelligence and vision-based intelligence to accurately detect and pick items in unstructured environments in warehouses and fulfillment centers, greatly improving picking speed and accuracy. Using machine vision and artificial intelligence, the item picker can determine the best gripping point for each item before the suction grip picks up the item and places it in a designated box. The system does not require any human supervision or information about the physical properties of the items it picks, is able to handle a wide range of items in dynamic and unstructured environments, and achieves high-precision picking with an accuracy of more than 99.5% with ABB vision systems, picking up to 1,400 unsorted items per hour. The product is currently suitable for a wide range of loads and applications and can be installed to one of three ABB robots – IRB1200, IRB 1300 and IRB 2600.

ABB has partnered with Microsoft to power industrial analytics with generative AI. According to the company's official website, on July 5, 2023, ABB and Microsoft collaborated to integrate generative AI capabilities into industrial digital solutions, and the cooperation aims to improve efficiency and sustainability by deploying Copilot capabilities to enable more intuitive user interaction with ABB Ability Genix industrial analytics and AI suite and its applications. ABB Ability Genix is a comprehensive, modular IIoT, analytics and AI platform. ABB will integrate generative AI into the Genix platform and applications through Azure OpenAI Service, including large language models (LLMs) such as GPT-4, for code, image and text generation. The new Genix Copilot app will enhance the user experience by providing intuitive functionality and streamlining the flow of connected data across processes and operations. Improve decision-making and increase productivity by providing real-time, actionable insights to industry executives, functional experts, and shop floor engineers. According to ABB, such insights have the potential to extend asset lifecycles by up to 20 percent and reduce unplanned downtime by up to 60 percent.

This article is for informational purposes only

Special report on the robot industry: the evolution trend of industrial robots from ABB

Sections viewed

Contact

working hours