2023 was a remarkable year for AI. Generative AI almost instantly emerged from an area of research to a working product even school kids know. Robotics installations growth continued to exceed expectations, and new innovation and products were introduced. What should we expect in 2024? A Lot.
The Dawn of Generative AI in Robotics
2023 marked a pivotal year for Generative AI, heralding a revolution in technology. For the first time, software and machines can autonomously generate content, including text, images, and videos, excelling with the internet's vast knowledge at their disposal.
This innovation is set to profoundly reshape our future. The way my children will approach writing, creativity, and work will be vastly different, diverging from traditional methods. The rapid integration of Generative AI into our lives is nothing short of astonishing.
Its impact on robotics, while certain, has an indeterminate timeline. We will see a lot of demos and experiments using GenAI in robotics, and it will be interesting to see which, and when, new products will hit the market. The potential is huge.
A Paradigm Shift in Robotics: from Hardware to Software
The robotics industry is undergoing a notable transformation, pivoting from a hardware-centric to a software-driven model. This change paves the way for a multitude of Generative AI applications. Although the robotics industry typically adopts new technologies slowly due to its inherent complexities, I am optimistic that Generative AI will catalyze significant industry changes within five years. Initially, we'll see basic implementations in robotic arms and mobile robots, facilitating control with commands like “generate code to move the robot along a specific path,” substantially simplifying the coding process.
A major turning point will occur when Generative AI can configure a robot for an entire task. Presently, manual coding for each task prolongs robot deployment. Imagine giving Generative AI a list of sensors, motors, and controllers, and asking it to control a robot for a complex task, such as “navigate this map and pick all the ripe tomatoes in this greenhouse.” The possibilities are immense. Note that achieving this level of sophistication will heavily depend on advancements in visual perception.
The Expansion of Mobile Robots Beyond Warehouses
Traditionally, commercial mobile robots have been confined to indoor environments, such as warehouses. These robots have evolved from isolated delivery units to key elements of cloud-based productivity ecosystems. The focus is now shifting to middleware and its role in coordinating mobile robots' tasks within larger Warehouse Management Systems (WMS) or Facility Management Systems (FMS).
Often, it seems that robots are limited to solving problems within their capabilities rather than being tailored to specific facility needs. Currently, robots are optimized for flat, indoor surfaces in stable environments with few changes, people, or machinery. They excel at repetitive tasks in predetermined locations. With the growing presence of robots in warehouses, a shift from traditional LiDAR sensors to vision-based technologies is likely. These sophisticated systems, powered by AI, will enable robots to navigate dynamic environments. This shift will prompt operators to automate previously untouched areas like loading docks and inter-warehouse spaces, introducing new capabilities such as forklift tracking and driving assistance—hallmarks of the next industrial IoT wave.
As robots advance with new perception technologies, a market for applications beyond warehouse walls emerges. We're observing an increasing demand for logistics automation in various settings, including hospitals, hotels, and retail, utilizing technologies akin to those in eCommerce robots. The restaurant robot sector, aiding waitstaff in food and dish transportation, is rapidly expanding. Agriculture is poised for significant automation advancements, illustrated by John Deere's ambitious plans to establish a fully autonomous farming system for row crops by 2030. In 2024, we anticipate a surge in robotic automation across industries such as agriculture, construction, logistics, and the food sector.
Humanoids Will Not Take Over The Workforce
The notion of robots usurping jobs is a longstanding misconception. There's a need for the robotics industry to better inform customers and end-users about the unique benefits of robotics. While humanoid robots won't supplant human workers, advancements in robotic arms, mobile robots, location aware human-level perception, and computer vision technologies will open up novel applications and uses over time. Robots represent an intricate tool created by mankind to boost productivity, often employed for tedious, dirty, or hazardous tasks.
When considering humanoids, we must differentiate between their capabilities:
Bipedal Mobility: Walking on two 'legs' offers distinct advantages, such as navigating rough terrain or climbing stairs. However, wheeled machines also have merits, such as rapid movement on flat surfaces. This comparison raises intriguing questions about the evolution of mobility in nature and robotics. Bipedal robots are swiftly advancing, with practical applications on the horizon.
Humanoid Structure: The typical humanoid form—one head, two eyes, a mouth, two hands, a torso—is shaped by biological constraints. It's debatable whether this is the most efficient design for a task-specific mobile, electrical machine. Imagine the utility of having additional eyes, arms, or a longer neck. Our design choices may reflect more of an attempt to replicate nature than to optimize functionality.
Attaining a human-level intelligent humanoid remains a distant goal. We hope that breakthroughs in Generative AI and our Visual Perception system will narrow the gap.
Investment in humanoid robots, such as the recently unveiled Digit robot, will proliferate for particular applications, propelling significant progress. Nonetheless, mainstream adoption of humanoids is unlikely in the near future.
Looking Forward: A Year of Innovation and Growth
As we kick off 2024 we look forward to an exhilarating year. Vision technologies, endless compute resources and new AI models had matured, setting the foundations for much needed innovation in robotics and automation. I can’t wait to see what’s coming.