Brain Corp: Automating the next generation of robots

About Brain Corp

Founded in 2009, Brain Corp is an AI software leader that powers the world's largest fleet of autonomous mobile robots operating in commercial indoor public spaces. Global manufacturing partners use Brain Corp's BrainOS AI platform and its cloud-connected autonomy service to successfully build, deploy, and support commercial robots at scale across industries and applications.

Industries: Retail & Consumer Goods
Location: San Diego, CA

Tell us your challenge. We're here to help.

Contact us

By migrating its AI software platform to Google Cloud, Brain Corp maximizes growth, improves robotic performance, accelerates development, and reduces costs.

Google Cloud results

  • Enables robotic communication in real time vs. every 60 seconds
  • Cuts robots' yearly infrastructure costs by 60%
  • Deploys software updates 6x faster
  • Accelerates new service development
  • Improves sales opportunities and customer retention

26,000+ public-facing AMRs just got smarter

Just over 100 years ago, horses were the main form of transportation. And although early innovators were selling and using cars and airplanes, few had them. Brain Corp envisions a similar trajectory for robotic development and adoption. Today, everyone knows about robots, few people have them—and soon they'll be ubiquitous.

Brain Corp's technologies currently power more than 26,000 autonomous robots, including the world's largest fleet of autonomous mobile robots (AMRs) operating in public spaces. Used by leading retailers as well as airports, malls, warehouses, hospitals, offices, and schools, AMRs include floor scrubbers, vacuums, and inventory scanners. And unlike robots that operate in manufacturing plants, these AMRs navigate a slew of dynamic challenges, such as people intentionally or unintentionally obstructing their paths.

To maximize ROI, Brain Corp focuses on what it does best: AI and ML software development. Instead of building turnkey robots, Brain Corp develops and manages its hybrid-cloud BrainOS AI software platform, and makes it available to original equipment manufacturers (OEMs), who use the platform to enable their new or existing machines to run autonomously.

Autonomy requires adaptability

After initially deploying BrainOS on AWS, Brain Corp decided to migrate its platform and engage Google Cloud as its innovation partner. John Gill, senior software engineering manager at Brain Corp, says, "Robots generate a lot of structured and unstructured data. We needed a cloud platform that can scale to store, process, and secure all that data in real time." Additionally, most of Brain Corp's partners and customers already use Google Cloud, and they're familiar and comfortable with its services, including those supporting storage and security.

Cloud-native services automate robotic processes

To improve its AI platform's serviceability and scalability, Brain Corp redesigned BrainOS's monolithic architecture, which had just one API, to a modern microservices design. "This was a rewrite from the ground up, with zero lift and shift and 99 percent new code," says Gill. Engineers saved time and simplified the effort required to meet rigorous scalability, safety, security, and data privacy requirements by using 28 Google Cloud services. Gill continues, "By leveraging native Google services, we get out-of-the-box coverage for many of our requirements and keep the size of our own teams down." BrainOS's key Google Cloud services include:

  • Google Kubernetes Engine Autopilot provides fully managed Kubernetes clusters with built-in security and autoscale capabilities.
  • Google Cloud Storage ingests and stores any kind of data sent by robots, including images and telemetry readings. Every robot has its own storage bucket, which eliminates the need for separate ingestion services.
  • Google Pub/Sub runs a real-time, event-driven messaging backbone between services. For instance, any time a robot writes data to Cloud Storage, Pub/Sub detects the event and facilitates the correct next steps. They likely include pulling the data out of Cloud Storage, processing it with Dataflow, and then storing the data in BigQuery or Firestore.
  • Google BigQuery serves as Brain Corp’s primary data lake for all robotic, operational, and business data, enabling its use for analysis and protecting it indefinitely.
  • Google Cloud Firestore stores documents used by mobile and robotic applications to deliver customer-facing experiences including application interfaces and visual maps.
  • Google Dataflow facilitates real-time data analytics that support customer insights and AI and ML algorithms.
  • Google Cloud Logging enables Brain Corp to analyze event log files to monitor robots' performance and identify development opportunities.
  • Google Cloud Identity Platform and Google Identity-Aware Proxy services authenticate users and govern their data access and use.

Anytime, anywhere robotic insights transform experiences

Because the redesigned BrainOS Cloud platform now communicates with robots in real time rather than syncing data every 60 seconds, Brain Corp was able to develop a mobile app that gives customers anytime, anywhere information about their robots. Gill says, "Our services in Google Cloud provide insights that enable us to improve robots' mobility in dynamic environments and customers' interactivity. Now when a customer pushes a robot's 'on' button, within seconds they can see real-time information about that robot on their mobile app." Brain Corp is also developing more sophisticated AMR insights such as an instant alert if a robot runs out of power, along with a map that depicts the robot's location.

Brain OS Architecture
Brain OS Architecture

Thousands of robots migrated in just three weeks

After Brain Corp established its Google Cloud design, it took six months to migrate all robots, even though engineers had to wait to install the software updates during each robot's idle time. "We migrated thousands of robots across the globe to Google Cloud in a way that ensured the robots didn’t exhibit any change to our customers," Gill says. "They were online when customers needed them to work."

By the end of the migration cycle, engineers had expedited their software processes so that they were able to update thousands of robots to run on Google Cloud in just three weeks.

Smarter processes cut costs by 60 percent

By migrating off AWS, Brain Corp has lowered the cost of robots. "With our legacy platform, we spent about 8–10 cents for every robot every day on infrastructure," explains Gill. "With our platform in Google Cloud, we spend 4–8 cents for each robot's daily infrastructure costs. This saving really adds up when you're talking about thousands of robots and counting." Operating a robot that costs 10 cents a day adds up to $36.50 a year. Multiplied by 20,000, that's $730,000. Operating a robot that costs 4 cents a day adds up to $14.60 a year. Multiplied by 20,000, that's $292,000, which is a 60 percent savings. "Our architecture in Google Cloud is scaling to support more workloads without driving the same level of exponential growth in costs." Gill says.

"Our architecture in Google Cloud is scaling to support more workloads without driving the same level of exponential growth in costs."

John Gill, Senior Software Engineering Manager, Brain Corp

Pushing software updates 6x faster

Engineers spend less time on infrastructure management and software updates now that they use cloud automation and a microservices-based platform design. "The automation, tooling, and deployment capabilities that we have in Google Cloud are just fantastic," Gill explains. "Once we have developed a new release, we can now deploy it into production in 10 or 15 minutes. In the past, deploying one release could take over an hour."

Converged data supports new partner and customer services

By consolidating its robotic and business data in Google Cloud, Brain Corp can offer new services, insights, and improved customer experiences. "We've only begun to scratch the surface of how we can use our data to drive better ROI and customer interactions," Gill says. "By integrating event data from our robots with data from our business and operational systems in BigQuery, our analytics team is driving insights that were not possible before." Additionally, Brain Corp can now give partners and customers a shard or pond of its data lake instead of aggregated datasets in a CSV file so that they can explore raw data. And unlike the previous platform, any number of people can pull and analyze data in Google Cloud without impacting application performance.

"By integrating event data from our robots with data from our business and operational systems in BigQuery, our analytics team is driving insights that were not possible before."

John Gill, Senior Software Engineering Manager, Brain Corp

Collaborative robotic development with Google

Recognizing Google's expertise in delivering AI technologies for the retail sector, Brain Corp partnered with Google to help shape its Store Vision AI product so that it would be ready to debut at Google Cloud Next. Gill says, "With this offering, we're going to help retail customers really monetize their data. The same robot that cleans customers' floors at night can also scan the shelves, enabling immediate inventory insights. If we were working with any other cloud provider on this Store Vision AI technology, we'd be two to three years behind where we are today."

"If we were working with any other cloud provider on this Store Vision AI technology, we'd be two to three years behind where we are today."

John Gill, Senior Software Engineering Manager, Brain Corp

Brain Corp is also developing a virtual tour service that enables managers to visually walk through an entire store without having to travel. Gill says, "Now that we're in the active adoption phase of a global robotic revolution, we're building consumer confidence and helping people understand that robots are really just fancy appliances. The amount of opportunity in this space is exciting, and we now have the cloud platform and partnerships we need to scale our robotic services to meet the exponential demand that’s already in motion."

For more information about Store Vision AI, watch this session from Google Cloud Next.

Tell us your challenge. We're here to help.

Contact us

About Brain Corp

Founded in 2009, Brain Corp is an AI software leader that powers the world's largest fleet of autonomous mobile robots operating in commercial indoor public spaces. Global manufacturing partners use Brain Corp's BrainOS AI platform and its cloud-connected autonomy service to successfully build, deploy, and support commercial robots at scale across industries and applications.

Industries: Retail & Consumer Goods
Location: San Diego, CA