Five years ago, the acquisition of Kiva created a market opportunity for a new wave of companies developing warehouse robots. While material handling equipment has always been a big part of distribution centers, this latest class of robots operates with an unprecedented level of autonomy powered by advanced sensors and software. Warehouse robots are a core focus of Schematic Ventures and, in this article, we hope to share our perspective.
Transportation & Manipulation
Warehouses, distribution centers, fulfillment centers and parcel sortation facilities are massive buildings with piles of boxes whizzing around at high speed. The purpose of each underlying system can be difficult to grasp at first glance. However each function, task or process can be lumped into two categories: manipulating an item or moving an item.
General manipulation tasks are complex, involve a high degree of uncertainty and can have multiple solutions. Examples include receiving an unknown quantity of boxes of unknown condition, choosing a configuration of boxes to stack onto a pallet or identifying the correct item in a bin of similar but different items. Each of these situations involves perception, comprehension, evaluation and formulation of a decision.
Transportation tasks have programmatic solutions with limited variability. New warehouse robots do leverage autonomy to optimize the solution but automation equipment providing transportation functionality without autonomy has existed for several decades (carousels, conveyor, AS/RS, etc).
A typical item’s journey through a warehouse follows this pattern:
- item is received, identified, recorded and routed for storage
- transportation resources move to-and-from a storage location
- item is processed, packaged and shipped out of the warehouse
From the perspective of a machine, manipulation tasks (steps 1 & 3) are difficult and transportation tasks (step 2) are simple. Simple is cheap and difficult is expensive in the world of robotics. Most transportation robots cost less than $25,000 while manipulation robot costs can exceed $100,000. The difference lies in the hardware requirements. Transportation robots oftentimes need only basic sensors for obstacle detection within an ordered, pre-mapped warehouse environment while manipulation robots utilize industrial arms with custom end-effectors guided by multiple, expensive sensors for precise perception.
Infrastructure Requirements of Transportation Robots
Given the expansive web of interconnected systems, highly automated warehouses are designed as holistic systems, involve massive capital outlay and are expected to last for many years. Storage infrastructure like shelves, rack or mezzanines are structural, inflexible investments that cannot be easily altered in an existing facility. Infrastructure dependence is a significant distinction that further separates this category into transportation systems (transportation robots with dedicated infrastructure like Kiva) and standalone transportation robots.
Transportation systems are best suited for new facilities and adopted at the design stage. Whereas transportation robots can be seamlessly integrated into existing warehouses and enhance existing processes without altering the facility design.
As in all supply chain decisions, flexibility comes at a cost; transportation robot queues and congestion are common in aisles designed for manual workers. This trade-off is further complicated given optionality considerations. The constantly shifting supply chain environment, evolving demand forecasts and internal company dynamics can make the sacrifice of speed for flexibility the right decision.
Manipulation Strategies
Manipulation of items has been described as the holy grail of robot design.
“It’s about accuracy. While robots are applauded for their repeatability, random bin picking requires accuracy in the face of chaos.
The robot has to locate a part in free space, in an unstructured environment where the parts keep shifting positions and orientations every time a part is removed from the bin.
That requires a delicate balance between robotic dexterity, machine vision, software, computing power to crunch all the data in real time and a grasping solution to extract the parts from the bin.” ( www.robots.org)
However the lack of a universal manipulation robot does not mean today’s manipulation tasks remain unsolvable. By simplifying the environment of manipulation, robots designed for that specific application can achieve high accuracy under constrained conditions. Other approaches include supplementing perception data through additional sensors (force, vibration, stress) to improve item identification or leveraging end-effectors with interesting material properties. Even with these innovative approaches, universal manipulation remains the toughest problem to automate in the warehouse.
Given the cost of an industrial arm and perception system, it is easier to justify the high capital expense by ensuring constant utilization of the arm in a stationary position. Minimal change to infrastructure is needed if the arm replaces a repetitive manual process conducted in the same place.
Alternatively, mobile manipulation robots provide simplicity by combining transportation and manipulation functions in one robot. While this means fewer types of robots in the warehouse, storage infrastructure may have to be altered for accurate, consistent picking and expensive arms are left idle during periods of transportation.
Warehouse Robot Taxonomy
We’ve distilled the above concepts into four categories:
- Transportation Robots
- Storage Systems
- Stationary Manipulation Robots
- Mobile Manipulation Robots
These categories align well with both the problems being solved and backgrounds of the teams tackling the problems. Incumbent automation companies (Schaefer, Dematic, Vanderlande, Intelligrated) have been focused on retro-fit opportunities that integrate seamlessly into existing customer warehouses or larger systems with massive revenue potential. Smaller robotics teams out of academia make up the majority of the manipulation companies, continuing work started in school on end-effectors and picking challenges. These start-ups are better positioned to develop scalable robot solutions with lower price points that would be impractical for large incumbents.
We’ve made two investments in warehouse robotics and see continued potential in this category. Our retail, distributor and integration partners are enthusiastic about scalable, autonomous solutions and, if you’re building a company somewhere in the above landscape, we’d be happy to connect you with those opportunities.
This is the first in a series about warehouse robots. In upcoming articles, we’ll touch on best strategies for building a warehouse robot company and how customers evaluate the business case when purchasing warehouse robots.
Thanks for reading and thanks to Alex for his help.