Robotics development has moved from companies or individual roboticists starting from scratch when coming up with a new robot idea. Instead, more companies are building on the work of other companies from the past 10 years, taking components, software or modules from them to quickly create the next big thing. In the case of autonomy for mobile robots, many have turned to RBR50 2019 honoree Brain Corp and its BrainOS.
The San Diego-based company specializes in developing intelligent, autonomous navigation systems for everyday machines. Founded in 2009, the company has earned acclaim for providing autonomy to commercial floor-cleaning vehicles, including a well-publicized deployment with Walmart. The company continues to develop its advanced machine learning and computing vision systems for the next generation of self-driving robots.
Robotics Business Review recently spoke with Phil Duffy, Vice President of Innovation at Brain Corp, about whether mobile robot autonomy has been mostly solved, the challenges of outdoor autonomous navigation, and why roboticists should avoid “starting from scratch” these days.
Is mobile autonomy solved?
Q: Adding autonomy to mobile vehicles, or adding better autonomy to mobile robots, has been at the forefront of the BrainOS technology. Have a majority of the problems with autonomy been solved for the most part, or are there additional challenges ahead?
Duffy: Really good question. Mobile robotic autonomy is solved on three levels:
- Ensuring safety in a public environment.
- Being aware of where the robot is in its environment and knowing where the robot is if the environment changes.
- Having the ability to navigate in complex, dynamic environments, i.e. avoiding people and obstacles.
Solving these three challenges and safely deploying fleets of robots in large and varied environments such as shopping malls, airports, big-box retail, offices, warehouses, and grocery stores continues to be a core strength for Brain Corp. Mobile robotic autonomy is an evolving science that depends on effectively and efficiently decoding edge cases in continually changing conditions and circumstances.
Today, and as a direct result of advancements in our robotic technology, we can resolve a greater number of edge cases in far less time.
Q: Brain has added autonomy to commercial cleaning vehicles, licensed the BrainOS technology to other mobile robot companies, and introduced a concept for a delivery robot, yet there seems to be other types of vehicles that could be “autonomized” – are you exploring other designs and types, or is there still a huge market for vehicles within those spaces (cleaning, delivery, retail, etc.)?
Duffy: Our current partners view Brain Corp’s transformative core robotic technology as extremely beneficial to expanding their businesses with new form factors within their market sectors. They know they can rely on Brain Corp to help them successfully produce, deploy, and support commercial robots powered by BrainOS across multiple industries and applications.
Robots are differentiated by their physical structure, user interface, and custom features. However, the majority of technical effort in creating these robots is dedicated to developing the foundational navigation system, which is designed to safely operate in diverse environments. BrainOS is a cloud-connected, foundational-level technology. In addition, BrainOS is application-agnostic, allowing robot manufacturers within a wide variety of industries to build their robots at a lower cost. Potential partners for BrainOS-enabled products include the delivery, security, healthcare, and inventory analysis industries.
Where to start with autonomy
Q: Other companies sometimes decide to create a new vehicle/robot with autonomy rather than taking an existing vehicle and adding autonomy – what are the advantages to the approach that BrainOS takes, compared with “starting from scratch”?
Duffy: Building a robot from the ground up is a complex, technical feat. Money, time and countless hours of R&D often goes into creating robotic technologies. Brain Corp eliminates the need for other robotic companies to reinvent the wheel.
By integrating BrainOS, we can save robotic developers approximately 300,000 R&D hours and $30 million in development costs, which is equivalent to a 75-person R&D team working for two years.
BrainOS can be added during the manufacturing phase to both existing product lines, and new product builds providing a quicker pathway to market, with functional quality products customers are already comfortable with. This allows robot builders to focus on their go-to-market product differentiation, greatly streamlining a path to profitability.
Q: What technology advances have allowed for innovation on the AI and Machine Learning space for the improvements in autonomy – hardware (Lidar), processing (AI chips), networking (edge and/or 4G/5G/Wi-Fi), cloud computing, better data sets, sensors, etc.?
Duffy: The explosive growth in the use of sensors in the automotive industry, combined with more advanced technologies used for industrial robotics, has greatly enabled mobile robot development and propelled the evolution of the autonomous mobile robot (AMR) space. As edge processing and chip and sensor technologies mature, we are running more complex algorithms that equate to lower cost systems.
The 5G low-latency promise may enable a new wave of cloud computing to support robotic technologies such as image recognition. However, core autonomous capabilities, such as safety assessments and navigation, should continue to be computed onboard to ensure high levels of operational safety.
Challenges of outdoor environments
Q: There still seems to be a big gap between indoor autonomy (many systems available) versus outdoor autonomy (not so much). What is the nature of outdoor environments that makes autonomy more challenging?
Duffy: One of the major factors to consider when developing outdoor applications is the unique safety and compliance requirements for each environment. One of the more mature markets for outdoor robotics is agriculture, but the use case is generally application-specific and often does not transfer to other outdoor applications.
For example, robots that operate on public roads need to operate in compliance with both federal and local government rules, along with any self-driving car regulations. Robots that operate in other public outdoor environments, such as last-mile delivery, also need to meet local and state regulations. These specific requirements increase the cost, time-to-market and customer adoption appetite for those types of applications.
Q: How can other robot companies benefit from adding BrainOS to their existing robots, or if they’re developing new ones? Do many roboticists still feel that “they could/should do it themselves?”
Duffy: The fundamental challenge other robot companies have experienced is the lack of available, foundational robotic software that can securely scale.
To achieve safe and effective robotic functionality, BrainOS includes the sensor driver hardware abstraction layer, which consists of application-specific firmware and simulation tools, a BrainOS security layer, core middleware libraries, plus other foundational robotic capabilities such as odometry, perception, localization, mapping and motion planning.
Brain Corp knows operations scalability is critical. To jump-start improved scalability and ensure success, an API framework with applications to manage the autonomy, manufacturing, and deployment functionalities is available. This provides other robotics manufacturers with the hard-won knowledge we have learned over the past 10 years. To further empower those partners, BrainOS comes complete with cloud services for fleet lifecycle management, logistics reporting and more.
Since the BrainOS architecture is hardware agnostic, it can be easily adapted across different form factors and drive systems. Each motorized, wheeled vehicle would be able to perform diverse, specialized functions in different environments such as inventory scanning, material handling, or floor cleaning.
Q: What does it mean to the company to be in the RBR50?
Duffy: Achieving a vision of a world where the lives of people are made safer, easier, more productive, and more fulfilling with the help of robots takes dedication, teamwork, and tremendous cross-industry support. For Brain Corp to be in the RBR50 means that it has earned the respect of industry analysts who see the value BrainOS has contributed to the robotics industry, as well as our OEM manufacturing partners and customers.