Case Study

Autonomous Mobile Robot platform for RFID data capture in Retail Inventory Management

Download PDF

Our Client

Zone24x7 is the Innovations Partner for multiple Tier 1 Retailers in the US where we conceptualize, design and implement technology proof of concepts, demos, prototypes to explore the capabilities of latest technologies and evaluate their usefulness for the client organization. Zone24x7 Robotics team design and develop custom robotic solutions for specific business needs.

Challenges

Retail industry has embraced RFID technology as a key linchpin in successful Omni-channel retail strategies. RFID provides excellent item level visibility, which is a key enabler of improved inventory accuracy. At the early stages of RFID deployment, large retailers used handheld RFID readers to perform cycle counts. However, as more products became RFID tagged, increased human effort and time was required to perform frequent cycle counts, which increased costs.

As the duration between two cycle counts increased to weeks and months, retailers realized the need for automated hands free cycle counting to increase the frequency of cycle counts without necessarily breaking the bank.

RFID fixed reader solutions were an option but they were expensive to deploy in large stores operated by big department store operators in US. It is in this backdrop that our client who was already experimenting with RFID decided to evaluate mobile surface robots as an option for hands free cycle counts.

Frequent cycle counts were expected to increase inventory accuracy and reduce excess inventory and help stores fulfill online orders efficiently and accurately. In addition, our client wanted to help associates locate items faster during order pickups in the sales floor. Hence, they also required a solution that could locate RFID tagged merchandize.

How We Helped

Zone24x7 designed AZIRO robotic platform is a mobile surface robot with autonomous navigation, which could perform automated inventory taking and cycle counting of RFID tagged merchandize for retailers. AZIRO can overcome the dual challenges of frequent cycle counts while offering a practical means to estimate the location of an RFID tagged merchandize in a large floor area.

Retail sales floors are complex and challenging for an autonomous vehicle owing to the dynamic nature of store layouts due to seasonal trends etc. In order to overcome this complexity, AZIRO is equipped with multiple 3D depth sensors, laser range finders, sonar sensors and bumper sensors. All sensors are fused together to minimize processing overhead and power consumption with a unique fusion framework, which works as a centralized processing “brain”. This facilitates real-time obstacle detection and avoidance during autonomous navigation. The design provides flexibility to add more sensors with minimal overhead on processing power.

The robot is only allowed to operate in the sales floor once the shop is closed for its customers. We developed a scheduling system, which could schedule the robots for inventory tasks. The robot wakes up before the daily inventory scan, performs a diagnostic check of its critical internal systems and leave its charging station on a specified time without any human intervention. Store can also define the order and priority of departments for inventory scan.

Once the daily work schedule is completed, AZIRO returns to its charging station and connect to its cloud hosted software backend to upload the captured data. AZIRO was integrated with a robot fleet management system, which allows remote intervention, control and troubleshooting of a robot. Hundreds of robots could be monitored through this system to ensure the smooth operation of a robot fleet.

Sales floor maps generated by AZIRO, provided an additional input to the store management in planning floor layouts and fixture positions.

Impact of Our Work

AZIRO was tested in two store locations in the United States performing daily cycle counts. During the testing period it assisted in significantly increasing inventory accuracy. Store associates who were performing order pickups for Buy Online Purchase In store (BOPIS) and Ship from Store (SFS) were provided with an application installed in their handheld devices.

Products that were difficult to locate in the sales floor were verified with AZIRO backend to determine the location of the product. This quickly became the most sort after feature among all AZIRO features as AZIRO generated product locations dramatically reduced time spent by store associates on locating “trouble items”.

Just prior to the peak week in 2017, AZIRO managed to reduce overall trouble items by more than 60% while the rest of the store network experienced increases in trouble items up to 30%. Store Managers also confirmed major savings due to reduced hours spent on locating items which were difficult to find.

After analyzing several months’ worth of data, AZIRO team produced heat maps of products which were slow moving yet occupied valuable shelf space in the sales floor.

Download PDF
zone24x7 Logo