Autonomous Mobile Robots Take a Direct Route to Optimization With AI

While the use of Artificial Intelligence (AI) continues to surge, its definition and how it’s applied often varies depending on application and industry sector. In the world of autonomous mobile robots (AMRs) for instance, AI takes the form of a system that collects data, then learns and adjusts as the data changes. Essentially, this application of AI in its basic form is the optimization of data and is typically found in manufacturing/production environments as opposed to huge warehousing operations.

For AMRs, AI helps optimize the flow of materials within a facility through the use of collected data, which is then incorporated into AMR fleet management software. AI-enabled AMR fleets allow facilities to replace big, bulky manually operated forklifts with agile, more efficient AMRs. AI will also pave the way for AMR use in more challenging applications such as those taking place outdoors, in agricultural settings or in freezing environments.

Fleet management software using AI optimizes the routes AMRs take along pre-configured paths. These routes are the structured paths that a facility wants AMRs to follow. The pathways are embedded with various nodes for charging, loading, unloading and other operations.

The optimization of an AMR fleet’s movements along pre-configured paths is how a facility coordinates the logistics or manages the fleet. AI-powered fleet management software directs AMR traffic throughout a facility, ensuring efficient flow and avoiding collisions. While this is relatively straightforward for two or three AMRs, it becomes quite complex as the number of AMRs in a fleet increase.

Consider a facility with a fleet of more than 200 robots. What is the most optimized way to use them, which robot goes where and does what task? At any point along their set paths, they can make decisions — right, left, back forward — based on collected data, so in addition to playing a role in how a robot goes from point A to point B, AI also optimizes the processes that are happening at points A and B.

AI-Powered Software

An example of AMR fleet management software with AI functionality is KUKA’s Mobile Robot expert System (KMReS). The software not only enables comprehensive fleet management of an entire AMR system, but it also regulates all fleet traffic and is able to automatically reschedule and redirect in the event of obstacles.

For supporting AMR integration, the easy and intuitive system is a no-code-platform that lets facilities configure settings using a curser as opposed to programming them. It’s done through flowcharts, and users create nodes of robot actions that are tied together in the flowchart that the software then executes. With the software, users create, manage, and edit workflows as well as monitor and manage containers the robots are handling. All this allows new or modified routes to be planned quickly and efficiently. For experts, there are still more advanced programming options available, making the software useable even in uncommon applications.

Beyond managing multiple AMRs along preconfigured paths, today’s fleet management software allows those AMRs to also navigate around unexpected obstacles in their paths. Similarly, as the use of AI in mobile robotics increases, the platforms will use advanced sensor technology to not only detect objects in their paths, but also identify them.

Essentially, an AMR is a piece of hardware that depends on numerous sensors, including 3D vision systems and cameras. In addition to general navigation, they could use these sensor arrays along with AI to detect whether an obstacle is a human or inanimate object such as a pallet. This, in turn, means that the better 3D vision system and camera technologies become, the more effective is their object identification and thus their navigation capabilities.

Robot Vision System

In addition to AI-powered software, 3D stereo cameras have had a huge impact on the advancement of robot vision system technology. They allow robots to recognize parts — not just their location, but also their orientation. The 3D stereo camera/vision system captures a part image and transfers it to software, which then uses the images to extract data representing viable parts that the robot can pick. From the image, the software rates which part is in the optimal pick position or relatively close to it, then sends decisions to the robot.

The KMP 1500P’s cameras can also read QR codes. This can be used to reach a higher level of precision — positing accuracy of +/- 5mm — which is often necessary at handoff points where the robot is picking up or dropping off materials. In QR code navigation, a Simultaneous Localized and Mapping (SLAM) map is used as a reference for setting up the paths in the software, and the QR codes are placed on the facility floor to use it for navigation. Why use QR codes?

Consider a plant where some sections of the facility have a frequently changing environment. Rather than having to add physical features to make the SLAM navigation work, those facilities can use the QR codes to navigate the robots in these areas.

The KMP 1500P camera system permits safe, autonomous transport of heavy loads in factories and logistics centers. With its agile drive system, the KMP 1500P can navigate complex and dynamic environments, adapt to changing requirements and optimize material flow. This provides agility and versatility in operations, ultimately helping businesses to respond quickly to evolving market demands and achieve higher productivity.

Advanced Wheel and Drive

The flexibility and maneuverability of AMRs would not be possible without the advent of advanced wheel and drive technology. Two such advancements include KUKA’s omnidirectional platform wheels and the diffDrive differential-drive technology. Featured on the KMP 1500P AMR, diffDrive uses two centrally located driving wheels that are opposite of each other and four caster-type wheels at each corner. The system allows the AMR to pivot and turn on a single spot.

The omnidirectional drive technology is based on the Mecanum wheel and provides full 360-degree freedom of movement for unlimited maneuverability. They are electric motor driven and typically consist of two rims and nine free-running rollers mounted at 45-degree angles that move independently of each other. This allows automated platforms to move not only forward and sideways, but also diagonally — basically any movement on a plane is possible without steering.

While the differential drive systems need to rotate the AMR/platform to change the direction of motion, the omnidirectional drive systems allow movements in any direction, without changing the orientation of the platform.

Also on the robot side, additional software comes into play — an operating system — that is necessary for the vehicle’s own navigation and for it to communicate with the fleet management software. The robot will also have software for safety and basic driving control. Robotic vision systems work in tandem with this software that processes a camera image, then directs the action of the robot based on that visual information.

While advanced vision systems give AMRs the power of “sight” so to speak, AI allows them to identify objects and optimizes how they navigate on a factory floor. Using collected data and AI, current AMR fleet management software more effectively controls the flow of materials within a facility. Such capability provides these facilities with a viable alternative to traditional material handling, i.e. forklifts, and opens the door to applying AMRs to a range of more challenging applications.

This article was written by Denise Strafford, Regional Head of Advanced Robotic Applications at KUKA Robotics (Sterling, MI). For more information, visit here  .

---