According to the research team, large robotic warehouses are increasingly becoming part of the supply chain in many industries, from e-commerce to automotive production, but getting hundreds of robots to and from their destinations efficiently is ‘no easy task.’
As navigating the robots is such a complex problem, the MIT researchers said that even the best path-finding algorithms struggle to keep up with the pace of e-commerce or manufacturing.
To tackle this, the researchers, who use AI to mitigate traffic congestion, applied similar ideas from that domain to the new warehouse model.
The deep-learning model encodes important information about the warehouse, including the robots, planned paths, tasks, and obstacles, and uses it to predict the best areas of the warehouse to decongest in order to improve overall efficiency.
Their technique divides the warehouse robots into groups, so smaller groups of robots can be decongested faster with traditional algorithms used to coordinate the robots.
The researchers tested the model in several simulated environments, including modelled on warehouses, some with random obstacles, and even maze-like settings that emulate building interiors.
By identifying more effective groups to decongest, the research team said their learning-based approach decongested the warehouse up to four times faster than strong, non-learning-based approaches. Even when they factored in the additional computational overhead of running the neural network, their approach still solved the problem three and a half times faster.
In a statement, lead author and assistant professor in civil and environmental engineering at MIT, Cathy Wu, said: “We devised a new neural network architecture that is actually suitable for real-time operations at the scale and complexity of these warehouses.
“It can encode hundreds of robots in terms of their trajectories, origins, destinations, and relationships with other robots, and it can do this in an efficient manner that reuses computation across groups of robots.”
Traditional search-based algorithms avoid potential crashes by keeping one robot on its course and replanning a trajectory for the other, but the researchers said that with so many robots and potential collisions, the problem quickly grows exponentially.
“Because the warehouse is operating online, the robots are replanned about every 100 milliseconds. That means that every second, a robot is replanned ten times. So, these operations need to be very fast,” said Wu.
In the future, the researchers said that they want to derive simple, rule-based insights from their neural model, since the decisions of the neural network can be opaque and difficult to interpret. Simpler, rule-based methods could also be easier to implement and maintain in actual robotic warehouse settings.
This research, supported by Amazon and the MIT Amazon Science Hub, can be accessed in full here.
Durable PVC less likely to shed microplastics
inching towards the bottomless rabbit-hole called ´demonstrable innocuity´