Industrial robots emerged at the dawn of Industry 3.0, moving towards computer control and automation, and have evolved over many years into specializations across various industries and processes. Robots were designed for mass production; they are typically solitary, performing specific tasks in relatively isolated situations. With the advent of Industry 4.0, cyber-physical systems, and the Internet of Things (IoT), some robots have evolved into collaborative robots. Collaborative robots interact with their environment, including humans and other robots, and enable flexible manufacturing and mass customization (Figure 1).
Figure 1: Traditional industrial robots (left) operate in isolation, while collaborative robots (right) are designed to interact with their environment, including people and other robots or machines. (Image credit: Omron)
The evolutionary path from robots to collaborative robots involves many adaptations: collaborative robots operate differently; they are programmed differently; they tend to be smaller, simpler, and in some cases, mobile; they are used in different processes compared to robots, and they must adhere to different safety standards. Collaborative robots generally do not compete with or replace robots; rather, they expand the opportunities for automating processes.
This article traces the evolution from robots to collaborative robots: comparing different operational methods of robots and collaborative robots; reviewing different programming approaches used in collaborative robots; discussing the use of artificial intelligence (AI), the Internet of Things (IoT), and other technologies to enable the mobility and human interaction of collaborative robots; detailing some applications where cobots excel, such as process finishing operations, quality control, logistics/material transportation, etc.; and examining extended safety standards for collaborative robots. Throughout, it paints a picture of future cyber-physical operations that integrate robots, collaborative robots, and humans to maximize productivity and quality while minimizing overall costs.
Collaborative robots can not only work alongside humans, but also move from one place to another (Figure 2). These characteristics are of great significance to collaborative robot programming, the location and time of their use, and safety requirements.
Figure 2: Collaborative robots can move from one place to another as needed for a specific task. (Image source: Omron)
Teaching Collaborative Robots
Industrial robots are programmed using languages such as C and C++. Collaborative robots have evolved to the point where they can be "taught" using various no-code tools, such as pendants, tablets, and even manually moving the robot arm from one point to another (Figure 3). Employing different teaching methods instead of traditional programming allows collaborative robots to learn new tasks more quickly, which is crucial when transitioning from one task to another. The time required to program industrial robots is economically significant due to its relatively long duration in high-production applications. On the other hand, collaborative robots need to learn new processes quickly to avoid lengthy and costly downtime. Machine operators can teach collaborative robots specific tasks without the need for specialized programmers. Tasks such as pick-and-place, including visual inspection of results, can be taught to collaborative robots in minutes.
Figure 3: Collaborative robots can be trained by moving their arms from one location to another. The operator's right hand rests on a high-resolution camera, which the collaborative robot uses to see where it is and what's in that location. (Image credit: Omron)
Artificial intelligence combined with machine vision can help improve the learning and functionality of collaborative robots. Intelligent collaborative robot vision systems offer a range of capabilities, such as object recognition and localization, barcode and totem interpretation, pattern matching, and color recognition. Vision systems can also guide collaborative robots from one location to another via gestures and teach them new processes. In other cases, robot operators can quickly and effectively teach collaborative robots using a drag-and-drop flowchart-based system on a tablet (Figure 4).
Figure 4: Intuitive drag-and-drop teaching/programming maximizes the productivity and flexibility of collaborative robots. (Image credit: Omron)
In addition to collaborating with humans, collaborative robots can also work with autonomous mobile robots (AMRs) to move from one task to another (Figure 5). AMRs are specialized collaborative robots that work alongside humans, collaborative robots, robots, and machines to perform tasks such as material handling with exceptional efficiency. Like material handling, moving a collaborative robot from one location to another is not a highly skilled activity, making it well-suited for AMR implementation. AMRs navigate from one location to another by combining onboard sensors and computing to understand their immediate environment. Wireless connections to centralized computing resources and sophisticated sensor networks throughout the facility help AMRs understand the location of obstacles along their planned routes and effectively navigate around fixed obstacles such as workstations, racks, and robots, as well as variable obstacles such as forklifts, other AMRs, and humans.
Figure 5: A collaborative robotic arm (top) can be picked up and moved to a new workstation by an autonomous mobile robot (bottom). (Image source: Omron)
What are the uses of collaborative robots?
The ability of collaborative robots to work alongside AMRs, humans, other robots, and machines opens up new opportunities for automation. Collaborative robots are used in a wide variety of industries and processes for large-scale customization, such as assembly operations, distribution, screw driving, machine maintenance, automation, pick and place, and so on, in equally broad industries, from automotive to food processing and semiconductor manufacturing (Figure 6).
Figure 6: Collaborative robots are highly flexible and can be used in a variety of applications. (Image source: Omron)
Repetitive or complex assembly tasks can be efficiently performed by collaborative robots working alongside humans. When paired with AMRs (Autonomous Mobile Robots), collaborative robots can improve the implementation of complex picking operations and the transport of materials to the work site. Once the materials are transported to the end of the production line, the collaborative robot can quickly sort the products for shipment. Using machine vision and artificial intelligence, collaborative robots can inspect, sort, and pick finished parts from the conveyor belt and place them into cartons. Collaborative robots can quickly adapt their behavior to new products and seasonal changes.
Collaborative robots are suitable for a variety of manufacturing processes, including (as previously mentioned) machine maintenance, screw driving, and dispensing. CNC machine tools, stamping and punching machines, various cutting machines, and injection molding stations are all examples of machine maintenance tasks, and collaborative robots can free people from repetitive and potentially hazardous activities. Screw-driven collaborative robots increase precision and consistent torque, resulting in higher quality assembly than manual assembly. Various materials, such as adhesives, sealants, paints, and other finishes, can be dispensed with high precision using collaborative robots. The end effectors of collaborative robots are interchangeable, allowing the robot to move from one task to another as needed (Figure 7).
Figure 7: Cobot end effectors can be easily switched to perform any task. This provides the flexibility to switch to different production requirements with minimal downtime. The first two end effectors include a high-resolution camera for an AI-based vision system. (Image credit: Omron)
Inspecting finished parts or products is another area where collaborative robots with machine vision excel. If the parts are complex, a thorough inspection may require high-resolution images from various angles, necessitating the coordination of multiple fixed cameras. Alternatively, a collaborative robot with a single camera can identify the part to be inspected and move around it accordingly, capturing all the images needed for a complete visual inspection.
Development of collaborative robot safety
Safety considerations for collaborative robots are also evolving. Compared to industrial robots, the safety requirements for collaborative robots are more complex. Teams of collaborative robots and humans can combine the repetitive performance of robots with the individual skills and dexterity of humans in a combine harvester. Collaborative robots (and robots in general) excel at tasks requiring precision, endurance, and strength, while humans excel at solving imprecise situations and variable problems. Combining these complementary skill sets presents challenges related to safe interaction between humans and collaborative robots.
Safety standards for industrial robots are typically based on excluding operators from the workspace while the robot is in operation. Cobot's safety considerations include interaction with humans. Cobot's speed, torque, and force limits define safety standards, including emergency stops and protective stops.
An emergency stop for the collaborative robots is initiated by the operator; it halts all movement of the robots and removes power. Recovery from an emergency stop requires a restart. A protective stop occurs automatically when a person enters the protective space around the collaborative robots (Figure 8). During the protective stop, the collaborative robots remain powered. Furthermore, during the protective stop, the collaborative robots' motion encoders are monitored for unexpected movement. If unexpected movement is detected, power is cut off.
Figure 8: The wrist joint safety space (blue box) around the collaborative robot can be rectangular or cylindrical, defining a no-go zone. If a person working next to the collaborative robot enters the no-go zone, the robot will initiate a protective stop. (Image source: Omron)
Some collaborative robots are designed with two operating speed settings: one for maximum performance and the other for maximum safety. In the performance setting, assuming no one will enter the robot's protected space, the robot will operate at its highest productivity speed. If someone enters the protected space, the collaborative robot will automatically switch to the human-robot collaboration setting, reducing speed, torque, and force to achieve maximum safety.
Regarding the safety of collaborative robots, several standards and guidelines are constantly evolving. ISO Technical Standard 15066:2016 and RIA Technical Report 15.606-2016 both describe four collaborative technologies for reducing risks to human workers: safety-level monitor stopping, hand guidance, speed and separation monitoring, and power limiting (PFL) systems. TS 15066 is normative, detailing the steps required for compliance. TS 15.606 is informative, providing information and methods applicable to standard compliance.
RIA TR R15.806-2018 describes a method for testing the forces applied by a PFL system. The sensor system needs to comply with standards related to velocity and separation monitoring. For both the PFL system and the safety-grade monitoring station, a restricted area is required for security protection.
ISO 13855:2010 establishes guidelines for protective measures regarding the speed at which collaborative robots approach specific parts of the human body. It provides a method for determining the minimum distance from the detection zone/restricted area or the activation of safety devices to the danger zone.
Summarize
Collaboration is a hallmark of Industry 4.0 and cyber-physical systems, and collaborative robots are key players driving higher levels of collaboration. Collaborative robots are constantly evolving, making them easier, safer, and more flexible to use. Advances in collaborative robot teaching tools and artificial intelligence are making their use more intuitive. The ever-evolving human-machine interfaces (HMIs) for collaborative robots are improving productivity and quality in mass-customized production. Collaborative robots are not replacing robots; they are expanding opportunities for automation, and the lines between robots, collaborative robots, and humans are becoming increasingly blurred. As collaborative robots become more like colleagues than industrial robots, safety standards for collaborative robots are expanding and becoming increasingly important to ensure the safe realization of the productivity promise of collaborative robot-human collaboration.
