AI and ML are helping warehouse robots better understand their environment and how to interact with it, helping them to operate more effectively
Robots continue to make their mark in the warehouse, working alongside humans, and as they evolve, so do their perception systems.
According to Exotec, the growth of AI and machine learning is playing an instrumental role in driving this improvement, as robots are continuing to adapt to their surroundings and fine-tune their operations.
Robotic perception is the ability of a robot to perceive, comprehend and reason about the environment in which it operates. Perception is the first step in robotic movement, enabling it to make correct decisions when complimenting or in partnership with humans in the warehouse.
Louis Esquerre-Pourtere, head of research and development at Exotec, says: “There are many key elements of robotic perception in the warehouse that can be challenging.
“Robots must be able to consider all possible environmental variations, such as lighting, temperature, dust, and many other unexpected parameters.
“These parameters must be considered in both the hardware design and software algorithm to ensure that the robot is compatible with efficient, performance, reliability and maintenance-free, 24/7 usage.”
Artificial intelligence and machine learning are used for automatic picking in the warehouse, playing a key role in helping robots learn how best to handle different articles.
Esquerre-Pourtere says: “As there is a lot of variation in terms of object, number, location, size and material, ML and AI are hugely important in helping robots determine the best course of action in the warehouse. AI also helps robotic arms identify picking points and trajectories.”
Perception systems in collaborative robotics cobots also play a significant role. Both robots and humans have dedicated strengths and having them working in partnership with one another means they can fulfil their roles without always needing to work at the same place and time.
Esquerre-Pourtere says: “In the warehouse, humans are good at adapting, making decisions and solving unexpected problems, often without any outside assistance.
“On the other hand, robots are good at doing things with minimal environmental knowledge or control. Yet, even with technologies such as ML or AI, they still take more time and are less predictable than humans.
“Nevertheless, with good processes in place, organisations can maximise the strengths of both. For example, by giving the bulk of the repetitive responsibilities to robots and giving humans the more complex and adaptive tasks, companies can build a much more efficient operation.
“Evolution in robotic perception allows robots to be better adapted to their environment, but it does not mean the environment should be ignored when developing a product. Safety should always be paramount, so this should be programmed into any warehouse robot as a priority.
“In 2025, perception will enable robots to automatically recover from unexpected situations, provide adapted modes of operation, and increase performance and reliability.”
