kirill_makarov - stock.adobe.com
Physical robots have been around for almost a hundred years, but not without their limitations. In the 1990s the idea of the collaborative robot, or cobot, emerged to help find ways to put robots in closer proximity to humans, but it is only through the inclusion of AI that robotics can continue to progress.
While the term robotics conjures up visions of hardware machines performing a wide range of tasks, the term robot is now used to describe any sort of software or hardware-based automation that can perform a task.
However, many of these software robotics systems are limited in their ability and are not able to communicate with other systems or robots to carry out these tasks. With the addition of machine learning, robots and cobots can improve their communication and handle even more complex tasks without the normal risk associated with simpler bots.
Cobots and software
Cobots are physical robots that are intentionally designed to operate in close quarters with humans. They are finding increasing use in a variety of different settings, performing pick-and-pack warehouse activities, delivery of goods, and a variety of assistive roles. Increasingly, we are seeing cobots in places as diverse as retail stores, museums, hotels, hospitals and even inside homes.
In this context, robotic process automation (RPA) refers to those software automations that perform repetitive user interface-based tasks that would otherwise be performed by a human, such as typing, clicking, swiping, copying and pasting, and a range of UI-based interactions.
But, if a form layout changes, or additional fields of information are required, these bots are not able to process and handle these exceptions and changes, causing them to fail and making them very brittle.
How AI and machine learning are working with robotics
What makes a robot powerful is an ability to think on its own. This is where artificial intelligence and robotics can come together. Companies are increasingly looking for robots to move past automation and tackle more complex and high-level tasks.
AI can help a robot do a lot of tasks, from successfully navigating their surroundings, to identifying objects around the robot or assisting humans with various tasks such as bricklaying, installing drywall or robotic-assisted surgeries.
Robots can benefit from AI and machine learning in different ways, and these AI-enabled capabilities include:
- Computer vision. AI and computer vision technologies can help robots to identify and recognize objects they encounter, help pick out details in objects and help with navigation and avoidance.
- AI-enabled manipulation and grasping. Long considered a difficult task for robots, AI is being used to help robots with grasping items. With the help of AI, a robot can reach out and grasp an object without the need for a human controller.
- AI-enhanced navigation and motion control. Through enhanced machine learning capabilities, robots gain increased autonomy, reducing the need for humans to plan and manage navigation paths and process flows. Machine learning and AI help a robot analyze its surroundings and help guide its movement, which enables the robot to avoid obstacles, or in the case of software processes, automatically maneuver around process exceptions or flow bottlenecks.
- Real-world perception and natural language processing. For robots to have some level of autonomy, they often need to be able to understand the world around them. That understanding comes from AI-enabled recognition and natural language processing. Machine learning has shown significant ability to help machines understand data and identify patterns so that it can act as needed.
In the past, researchers have long thought about how to apply artificial intelligence to robotics but ran into limitations of computational power, data constraints and funding. Many of those limitations are no longer in place, and as such, we now may be entering a golden age of robotics. With the help of machine learning, robots are becoming more responsive, more collaborative, and integrated into other systems.
Likewise, many of the RPA vendors are adding intelligent process automation to their bots to help increase their usefulness. As such, they are looking at AI technologies such as NLP or computer vision to help make these bots more intelligent. Bots that leverage machine learning and adapt to new information and data can be considered intelligent tools that can significantly impact and increase the tasks performed rather than just bots.
Growth of robotics
The use of robots in many industries is becoming increasingly common. These robots can either be physical robots or software bots. It is estimated that there will be 3 million industrial robots in operation during 2020. Furthermore, Gartner projected that RPA software spending was over $1.3 billion in 2019. As such, the need and desire for bots of all sorts is seemingly only to going to increase.
Examples of AI powered robotics include: robotic surgery tools that are able to assist surgeons, law enforcement bomb robots that are able to navigate into dangerous terrain to minimize human injury and casualty, and food and package sorting robots that are able to sense different materials and properly pick and sort the objects.
With the use cases seemingly limitless and cutting across many sectors, there is much innovation still to be had and the robotics industry isn't going away anytime soon. Many companies are finding increasing value, efficiency and accuracy from bringing robots into their various operations. This stems from the proof of ROI in the industry and, as people continue to feel more comfortable working with robots, companies will continue to invest in the technology. The addition of artificial intelligence into robotics its making them more useful than ever before.