I sat down recently with CAJA Robotics co-founders Ilan Cohen (CEO and Chairman) and Hanna Yanovsky (EVP and CRO) to learn more about the company, their solution and the future of their business.
Israeli based CAJA (pronounced “caha”) Robotics was started in 2014, and has completely bootstrapped the company up to this point in its history. The company currently has 55 employees and has recently opened a US subsidiary to expand their business into North America. CAJA manufactures autonomous mobile robots (AMR) to automate warehouses.
Mobile Robot Guide (MRG): Tell us about your solution.
Hannah Yanovsky: “Our solution is designed to assist in warehouse applications where humans are walking kilometers per day and lifting heavy boxes on and off shelves all day. Our solution is differentiated from other collaborative robotic solutions and addresses the major pain points in a different manner from our competitors.
“The CAJA solution operates in a ‘sterile’ warehouse environment – where only the robots are in the workspace doing the fulfillment. Human pickers pull products for customer orders from the boxes/totes that are delivered to the pick stations. They don’t enter the workspace of the robots.
“We’ve designed a solution that can be deployed without any changes to the existing warehouse layout. This capability allows the throughput to be higher when compared to collaborative robotic solutions. Unlike collaborative robotic solutions, our solution uses all of the existing infrastructure as if it was in manual mode. We can acquire and move any cardboard box or plastic tote (up to 30 Kg).
“This market is so dynamic, our customers are demanding a solution that can evolve and that can be adapted to the changing needs of the customer. This is one the largest pain points for our customers. They can also deploy the solution gradually by automating one part of the warehouse while keeping other parts (of the warehouse) in manual mode. Our customers can also switch back to manual (as needed) or redefine the automated parts of the warehouse. Everything is managed from the cloud.
MRG: “How is your solution different from other automated warehouse solutions?
Ilan Cohen: “We did a lot of voice of the customer the last couple of months. Most of them say that our solution is the only solution (on the market) that resolves their needs. Some of the uniqueness is how we work from the cloud, how we work with two different kinds of robots and how we work in parallel – these are the advantages of our solution. For the customers, this is a positive impact on ROI.
”We’re hearing from many of our customers that they see that collaborative robots (and collaborative use cases) are good only for a limited percentage of the warehouse while the majority of use cases are ideal for our solution.
MRG: “Would you position CAJA Robotics as an AS/RS – given that there are no humans in the robot workspace?
COHEN: “By operating in a ‘sterile’ environment, we can operate the robots at a much faster speed and we don’t need to go through the full level of functional safety and (collaborative robotic) regulations, which is more and more difficult to obtain.
“All of our controls engine and advantages of software is in the cloud, including traffic management, warehouse management, optimization, synchronization. Everything is done in the cloud and this allows use to (continuously) improve the solution.
MRG: “By your definition, a collaborative robot is one that operates in the same workspace as a human?
COHEN: “Everything inside the warehouse is done by the robots. Human pickers work at pick stations outside of the warehouse floor. This allows the robots to optimize their operations and deliver the right products to the picking stations in the most efficient manner.
YANOVSKY: “Compared to other sterile solutions, our solution is at the same degrees of freedom like collaborative robots in terms of what it utilizes in the existing infrastructure. We benefit from the best of both worlds: sterile solutions can have higher throughput than collaborative solutions and CAJA can operate at a lower cost than collaborative solutions.
MRG: “You currently have two vehicle configurations?
COHEN: “We currently have a tall version of the robot and shorter version of the robot. We do collaboration between the tall robot and the small robot to hand off boxes. The tall robots can work the aisles, while the short robots deliver boxes to the picking stations. The tall robot can reach up to 3.4m (higher than a human) and acquire a box (up to 30Kg).
MRG: “Can your system be deployed into an existing warehouse, without any reconfiguration?
YANOVSKY: “Yes. The system can be deployed into any existing shelf structure as long as the box or tote is 30 Kg or less and is able to be slid on and off the shelf.
MRG: “Do you use SLAM for localization or do you use fiducials for localization?
COHEN: “We have a lot of vision and other sensors on the robots for safety. All of the traffic management is done in the cloud so we don’t need to keep a mapping navigation on the robots. We put a few fiducial stickers on the floor for individual robot localization. In the future we can add navigation by natural feature mapping, but it’s currently not our priority.
MRG: “Are you selling your solution using Robots-as-a-Service business model?
YANOVSKY: “We are open to such a business model. But so far the large customers are asking for a CAPEX purchase of the robots. A significant part the solution is software, the brain of the solution is the software, not the robot. Everything is managed in the cloud, including the machine learning, so we price the solution based on the overall value of the solution.
MRG: “Do you take all of the real time information into the cloud?
COHEN: “We can take a lot of real time WMS input data (into the cloud) to coordinate the robots with the WMS. Our software can optimize the warehouse inventory and coordinate with the WMS to do things like consolidate inventory. Our software keeps track of real time inventory. The system can optimize inventory locations to improve fulfillment based on predicted demand.
MRG: “Anything else important that we didn’t cover?
YANOVSKY: “Our 3PL customer is reporting 50% optimization in space, compared to what they had before. The system has the ability to consolidate and optimize inventory, including put away.
”We have two types of picking stations serial (synchronized) and non-serial (non-synchronized). With a serial picking station, the robot comes with the box to the picking station, the picker takes one product, and the robot returns the box. In this case, the robots queue up with the product totes.
”in a non-serial picking station, the tall robots can organize boxes into a put wall and then replace then as necessary to expedite the picking process. A put wall can also be used to queue high frequency products near the picking station.
“Our solution can also operate on a mezzanine, something that most collaborative robots can not do. This is a differentiator.