Operating a fleet of robots presents many challenges. The real world always has unexpected surprises and just because a robot works correctly in a lab doesn’t mean it will handle real world situations as intended. Even without an edge case occurring, the expected time to complete a mission may vary between locations or even time of day. A slight change in a given scenario may justify a different response, which robot operators can ideally account for and program ahead of time.
Building on the recent successful launch of Configuration as Code, we’ve developed a robust solution for users to programmatically define what we call Advanced Incidents, giving InOrbit users granular control of how a robot identifies and responds to any incident.
In practice, that may mean something as simple as specific instructions for a mislocalized AMR, or a unique response to more complex states. With Advanced Incidents, a robot’s state can be defined as broadly or narrowly as desired using any combination of dynamic values from data sources and user-defined tags. Unique actions can be then assigned as a response to any of these states for any subset of a robot fleet. These actions can be automatic or manual, which allows for even more user input.
This means that configuration can be applied broadly to an entire fleet while unique edge cases are handled in the exact way that our users anticipate. Parameters such as location, hardware, status, and data sources can be chained together to define a unique state that InOrbit uses as a trigger for responses.
Why is this so important for those managing robot deployments as they scale?
InOrbit’s Head of Product, John Simmons took some time to share his expertise and insights on InOrbit’s new capabilities. “The real world is complex and full of different environmental challenges for autonomous robots. In many cases, these challenges are not described by a single factor but require a combination of multiple conditions to effectively identify when an important event has occurred. These conditions range from the obvious, such as a robot that isn’t charging but also isn’t doing any work, to very advanced such as an alert that triggers when a robot is navigating more slowly than it should or when a zone in a warehouse is at its robot occupancy capacity.”
“InOrbit’s Advanced Incidents are designed to give robot operators the power to describe these scenarios using any number of data sources, spatial calculations, time ranges, and more. Even more importantly, Advanced Incidents require no changes to the robot software itself, decreasing the time required to implement new incident monitoring from weeks or even months to minutes.”
The right tools
A cornerstone of effective RobOps is observability. That’s why we continue to develop the best tools to intimately understand the challenges robots face in the real world. More than ever, we now have the solutions in place to truly address those challenges.
With Configuration as Code, we’ve allowed engineers to streamline their development workflow with software automation best practices to fit their specific needs. Directly speaking to the most common pain points that hinder and delay effective operational solutions including managing infrastructure as code, source-controlled configurations, Continuous Integration / Continuous Delivery (CI/CID) pipelines and more.
Advanced Incidents builds on this by providing the next step in the journey to programmatically resolve autonomy exceptions completely for every company using InOrbit tools to manage and operate their robots.
Giving operators more power
These tools are the building blocks in a framework of strong RobOps that supports operators of autonomous fleets. Advanced Incidents empowers our users with extremely flexible capabilities.
As John mentioned above, this is done without any changes to a robot’s software. Advanced Incidents are designed to facilitate the quick implementation of new incident monitoring. Perhaps most importantly, since these configurations can span robots from different developers, operators managing heterogeneous fleets can work efficiently in increasingly common real world deployments.