Key Trends in Industrial Robotics Underlie Diversification, Increased Sophistication, and National Competition

Subscribe To Download This Insight

1Q 2019 | IN-5351

Discrete, and to a lesser extent, process manufacturing is where robotics found their worth during the last 50 years. Given the established nature of this market, venture funds and commentary have been focused on the deployment of robotics outside of manufacturing environments; toward other heavy industries, logistics, retail, hospitality, and consumer products. But going forward, a significant amount of revenue related to robotics will remain focused in manufacturing, and the use of robots in this space will become far more expansive and sophisticated.

Registered users can unlock up to five pieces of premium content each month.

Log in or register to unlock this Insight.

 

Areas of Growth

NEWS


Discrete, and to a lesser extent, process manufacturing is where robotics found their worth during the last 50 years. Given the established nature of this market, venture funds and commentary have been focused on the deployment of robotics outside of manufacturing environments; toward other heavy industries, logistics, retail, hospitality, and consumer products. But going forward, a significant amount of revenue related to robotics will remain focused in manufacturing, and the use of robots in this space will become far more expansive and sophisticated.

Industrial robotics is not just growing in a general sense, but is deployed at greater volume in previously underrepresented verticals. Industrial robots were previously focused on high value-added discrete manufacturing sectors, typified by the automotive sector. This is largely because robots are capital-intensive devices, and this expense is compounded by the cost of maintenance, repair, and operations. As a result, high value-added products like cars and aircraft are a much more likely market to make up cost than with consumer products and low value-added apparel.

Industrial Robot

Slowly, this is changing, and over the last decade, industrial robot use has become diversified across industries. Due to advances in motion control, machine vision, precision, and dexterity, they can be used to assemble, sort, and tend smaller products like smartphones and electronics.

The following chart showcases the annual Compound Annual Growth Rate (CAGR) for industrial robot shipments from 2017 to 2027. There is limited growth in process manufacturing involving primary metals like steel and chemical products, and, expectedly, there is strong growth in high value-added industries like automotive, electronics, and machinery. Even low value-added sectors like textiles will see strong growth, due to advances in motion control and precision through the developments of fixed robotic solutions, such as those produced by SoftWear Automation.

CAGR Industrial Robots

Most of the growth is due to the immense demand stemming from China’s breakneck growth in manufacturing. China is the largest manufacturing nation, surpassing the United States in 2010. Given that China was accountable for less than 3% of the global manufacturing output in 1990, this is a remarkable achievement.

China’s demand for robots stems from several drivers. Wages are increasing, making the country less competitive as an offshoring opportunity, and it is receiving significant competition from developing manufacturers like Thailand, which are directing capital investment toward robotics. China also needs to strengthen productivity growth if it is to continue taking market share in high value-added exports from developed economies like the United States, Japan, and Germany. In reference to the United States, the period of accelerated offshoring (between 2000 and 2007) is remembered disparagingly, and there is increased emphasis from political and business leaders to strengthen American competitiveness in high-tech exports.

Western nations remain well behind their Asian counterparts. A recent study from the Information Technology and Innovation Foundation (ITIF) showed that, when accounting for wages, countries like China, the Republic of Korea, and Thailand are forging ahead with robot shipments, while Europe and the United States are seeing far less uptake. The tilt toward East Asia in Robotics is resulting in the diversification in industrial robot use, as the global supply chain for electronics and machinery are more heavily centralized in Asia than the automotive or aerospace sectors. This is, in turn, being balanced by the slow and gradual process of reshoring of manufacturing operations to the United States and Western Europe.

High Mix, Low Volume

IMPACT


Low-mix, high-volume manufacturing is the norm generally, and has been the standard environment for industrial robotics. Unfortunately for many manufacturers in highly developed economies, such as the United States, competitors like China are much better placed to succeed at this, due to wage competitiveness, and more importantly, because western nations like the United States face a severe talent shortage. As a result, high-wage environments are trending toward high-mix, low-volume manufacturing.

The market advantages of high-mix, low-volume production is clear: tailoring to specific customer demand, improved responsiveness, and lower inventory requirements for finished goods. However, this production scheme has traditionally yielded lower quality output because of customization and it is more susceptible to downtime.

For high-mix, low-volume manufacturing to succeed, there must be built-in flexibility in the environment, and machines have to be readily reprogrammable and easy to use for workers. Diagnostics and analytics also have to be optimized to reduce downtime and provide insight on how to customize products in an efficient manner. The following key trends are among those enabling robots to help companies transition to this more demanding type of operation.

Deep Learning (DL)-Based Machine Vision: Machine vision is used in many applications for manufacturing, including counting, visual stock keeping, inventory management (from raw materials to finished goods), quality control and product inspection (e.g., detecting defects or ensuring standards are met), robotics (industrial and cobots), predictive maintenance, and infrared heatmap. Applying DL techniques to machine vision significantly improves the ability of robots to perform complex tasks like object classification and improves motion control capacity in line with robot control-focused Software-Defined Kits (SDKs) like Energid’s Actin system.

By using DL models, the accuracy of the machine vision system improves as more and more data are captured and the model is further trained based on those data. As compared to conventional machine vision models, DL-based machine vision models can be deployed by users without significant coding experience, as the model undergoes self-learning based on data gathered. In addition, the models are capable of performing additional functions, such as defect classification and predictive analytics, and determining the ideal parameters based on big data. This leads to improved analytics, because machine vision requires high-quality datasets. In many cases, end users are not able to provide this. Machine vision companies often cite a lack of quality data to train models as one of the largest challenges they face. DL machine vision would typically require at least 1,000 images to train a commercially robust model.

Collaborate Robots (Cobots) and Autonomous Mobile Robots (AMRs):Cobots were in the news for the wrong reasons in late 2018, with closure of Rethink Robotics, but with the market leader, Universal Robots, subsuming much of the leftover resources, a large CAGR for a small market is likely during the next decade. Cobots are struggling to find their unique role separate from industrial robots, and many are being used for standard tasks (pick and place, tending), but at a more affordable price. Advances in machine vision and motion control mean that these systems will eventually become more capable, being able to operate swiftly alongside humans with greater payload capacity. Improvements in haptics, control, and command will make deployments swifter and easier, while the revolution in robot mobility (provided by sensors and machine vision) will eventually allow AMRs to move and manipulate objects across workstations without human overseers. This is currently an aspirational concept with limited deployments, but is already happening in warehouse logistics via companies like Invia and IAM Robotics.

How to Ride the Transition

RECOMMENDATIONS


Industrial robots are slowly transitioning from a capital-intensive system of fixed assets that are unwieldy and require highly structured operational integration designed for high-volume, high value-added industries, to a more flexible technology that can be applied across both high value-added and lower value-added sectors in environments that are designed for high variation and low volume, and vice versa.

But this is not a time for celebration. Capital investment is not advancing at a sufficient pace in the west to catch up to the Asian markets, and no single Original Equipment Manufacturer (OEM) or service provider can provide a comprehensive solution. System integrators and service providers that can simplify the process of introducing industrial robots stand to gain, but this is hampered by the lack of familiarity in the industry about integrating service models within a manufacturing operation.

Companies that will do well will be highly-focused, third-party providers that can offer a plug-and-play solution that is platform-agnostic and allows the end users and system integrators to get on with the challenge of incorporating industrial robots across the operation. Examples of this include proprietary Operating System (OS) providers for robot navigation like Brain Corp, Seegrid, and Vecna Robotics, and platform-agnostic analytics providers like Senseye, SpaceTime, and SparkCognition. The challenge these third-party technology enablers face is they are dependent on the interoperability of systems provided by OEMs and the infrastructure of end users, which is often sub-par. These structural issues could be theoretically solved by an effective system integrator, but the challenge is monumental. More likely, system interoperability and factory infrastructure will progress in a decentralized manner, making flexibility the central tenet of anyone involved in smart manufacturing.