This article includes everything you need to know about autonomous mobile robots and their use.
You will learn:
An autonomous mobile robot (AMR) is a self-propelled self-powered mechanism designed to perform repetitive tasks or organizational functions using an internal guidance system. They are able to navigate their workspace using sophisticated software and mapping technology that makes it possible for them to “SEE” and observe their environment to complete a variety of tasks. Sensors, artificial intelligence, machine learning, and computer programs make it possible for an AMR to interpret the position of obstacles and avoid collisions.
The array of sensors on an AMR are constantly scanning the area around it to detect potential danger or obstacles. Once a problem is detected, an AMR will automatically plan a route around it in the most efficient way. The initial process for implementing AMRs involves providing them with mapping technology. Visual simultaneous localization and mapping (SLAM) technology makes it possible for an AMR to make navigation decisions based on its surroundings.
AMRs are often said to be able to “SEE” their environment. This aspect of their operation is a reference to the use of light detection and ranging (LiDAR) that uses pulsed laser based sensors to give AMRs the ability to measure distances. LiDAR serves as the eyes of an AMR and gives it a sense of its surroundings and location.
Out of necessity, the order fulfillment industry has had to find new technologically advanced methods to improve their distribution operations. The concepts of forklifts and manual picking processes lack the efficiency and speed required by modern customers. The response to this increasing demand for automated technologies has led to several innovations that operate through the use of computer programmed robotic devices.
The initial answer to the needs of automation have been automated guided vehicles (AGVs) that are programmed to follow tape, wires, reflectors, or other guidance methods to reach their destinations. Although AGVs have significantly improved order fulfillment, they have been limited by their need to be connected to some form of guidance mechanism and are unable to adjust their trajectory when confronted with an obstacle.
With the rapid advance of artificial intelligence (AI) and various forms of computer software, it has been possible for material handling companies to develop autonomous mobile robots that can move about a facility without the need of wires, tape, or guiding mechanisms. These highly advanced technical tools can be programmed to complete multiple tasks while avoiding personnel and potential obstacles.
Although AMRs are similar to AGVs, they differ in their amount of flexibility and autonomy. AMRs are capable of creating their own routes and finding the most efficient way to achieve their tasks. The effectiveness of AMRs makes processes and workflow more efficient and productive compared to traditional manual methods.
The implementation of robotic inventory systems is designed to take care of simple tasks that waste the time of personnel. Efforts to this effect have been reached and investigated for years and led to the implementation of robotic arms, conveyors, and quality checkers that have radically improved productivity and the quality of products.
A normal warehouse operation involves storing products, picking products, organizing products, and loading trucks or supplying production operations. In the past, these jobs have been completed by workers and a well organized set of racking. As technology has advanced, more efficient methods have been discovered that eliminate the need for manual picking and fulfillment. At the center of these efforts have been AGVs and AMRs.
Transporting inventory or product from one location to another is a very basic function and one designed for the use of AMRs. Order picking is a very expensive task due to the amount of time and personnel that is required to complete it. A worker walking from place to place in a facility takes up a major portion of their work schedule. Using AMRs as part of picking strategies reduces travel time by having the AMR get the item and bring it to the picker. An AMR picker carries carts, items, and products between workers and workstations. The worker selects the item they require without having to leave their workstation. The AMR can move on to its next task or other workers.
In zone picking, an AMR goes to the zone where the item to be retrieved is located. A worker in that zone is guided by an augmented vision picklist or pick to light provided by the AMR. Once the item is loaded, the AMR moves on to the next zone or to shipping. Additionally, AMRs can be equipped with carts that can be moved between zones to be filled for packing, processing, and shipping.
Autonomous mobile robots are a valuable aspect of the sortation processes. The different uses of AMRs for sortation are differentiated by the varieties of handling technologies. Sortation involves the use of tilt trays, belt systems, and conveyor rollers and includes a fleet of AMRs with tilt trays. Sortation AMRs are collaborative robots that work with workers and chutes used for order positioning and locating.
AMR systems are capable of completing primary and secondary sortation processes due their flexibility and programming. Primary functions include receiving operations and decanting them into bins or totes. Secondary functions are organizing products for shipping to customers. In each case, sortation AMRs are responsible for delivering a package to the correct location.
A popular type of sortation AMR are ones with tilt trays that work collaboratively with workers and chutes for order positioning and locating. A camera reads a barcode to identify the item to be shipped and sends it to the proper chute. Once the AMR reaches the chute, the tray tilts and sends the picked item into a container. An AMR takes the collected order to the shipping department.
Aside from their use for shipping, sortation AMRs can also be used for consolidating and organizing returns. Items are collected by an operator who enters the item number and quantity, which the AMR uses to take the collected items to their location. Return AMRs have short term use and are programmed for other tasks once a return is completed.
One of the most challenging and demanding activities for manufacturing and distribution is accurate control of its inventory. Over the years, the method for keeping an accurate inventory was to do an inventory count every three months, six months, or once a year. At the completion of the count, the value of the on-hand inventory was or is adjusted up or down depending on shrinkage or growth. In either case, a company has to make adjustments to its inventory cost.
A modern technological solution for keeping an accurate inventory is an inventory tracking AMR. Using computer vision and analytics, inventory scanning AMRs collect real time data regarding on shelf inventory. An AMR system directs receiving, shelving, replenishing, and picking. It completes each of these functions using checking methods built into its system to identify errors and irregularities and manages cycle times to catch problems before they become an issue.
A sophisticated set of cameras mounted on an AMR reads item bar codes and records them for future reference. With a few glitches, an AMR has an accurate record of every item in a warehouse. Additionally, AMRs can be programmed to pick items from a production line such that it is possible to automatically transfer completed products from assembly to shipping or storage.
An AMR inventory scanner can be periodically programmed to make a cycle count of stored products and parts in order to ensure an accurate on hand count. The process makes it possible to catch errors, inconsistencies, and irregularities. The use of AMR technology prevents production downtime due to part shortages, helps with customer service, and prevents capital loss due to overstocks.
Collaborative AMRs are used to optimize logistics and work closely with human workers to assist in completing a variety of functions. An essential element of collaboration AMRs are safety features due to their close proximity to humans. Although it may be presumed that fences and barriers would be necessary, a more accurate representation is the implementation of tested safety programming.
The definition of human AMR collaboration takes several forms from no interaction with humans to sharing a workspace and performing tasks in a team concept. When humans and AMRs work together, the AMR delivers parts or items and removes completed items. Each task is performed using instructions from the human. These collaborations can become very complex with the AMR following a worker as they place finished products on a shelf and retrieve items for assembly.
One of the features of collaborative AMRs is their ability to perform multiple functions such as put away, pick, count, replenish, and sort, all of which can be built into their programming. Each of these applications can be performed individually or in unison depending on the design and programming of the AMR. The key to collaborative AMRs is their ability to eliminate walking time or walking distances, which keeps workers engaged on their assigned tasks.
The optimal performance of a cobot is dependent on the assistance of its human partner who gives it instructions. This is unlike other AMRs that move freely about their environment and receive instructions through their programming. A cobot is an assistant that helps its human partner in accordance with how applications develop.
Storage picking AMRs are scalable and require several adjustments to their environment to work properly. The main feature that may limit their use is the installation of special racking to which the AMR is attached for the picking process. Storage picking AMRs are specially designed to scale up and down the racking as they pick items off the shelves.
The dimensions, location, and design of the racking has to be in accordance with the needs of the AMR. The use of a scalable storage picking AMR requires a group of factors that are exclusively related to the site where the AMR will be installed. The number of AMRs varies from site to site with some sites having as few as five to ten AMRs while other sites may have 50 to 75 AMRs. This factor is integral to the implementation of a scalable storage picking AMR since it necessitates special adjustments in order to work appropriately, which will require changes to the use of other AMRs in a facility.
An interesting aspect of AMR use is in the hotel and restaurant industries where simple services are being performed without the use of personnel. There are various redundant tasks that hotel personnel perform that can be completed using an AMR and include scrubbing and vacuuming floors, delivering food, and picking up trash. This particular use of AMRs is in its infant stage and may improve with the implementation of future innovations.
Autonomous mobile robot forklifts are used in the same capacity as operator driven forklifts. They are capable of performing a wide range of functions efficiently and flexibly with little human interaction. Unlike traditional forklifts with propane or electric motors and a driver, AMR forklifts take up less room and operate noiselessly using a 3D detection mechanism.
A huge benefit of AMR forklifts is their ability to be immediately integrated into the functions of a warehouse by programming their software to fit the conditions of the warehouse and its environment. A necessary part of their function, and one of their main advantages, is their ability to quickly adapt to changes in their surroundings.
A common part of forklift operation is constant change in routes and picking requirements. In the middle of a picking, further instructions may require a change of route to a new location. Forklift AMRs easily adapt and change to meet any new programming and instructions. As with other forms of AMRs, forklift AMRs read their environment and use their sensors to guide their path. When given new instructions, they choose the most efficient way to a location and immediately initiate it.
Autonomous mobile robots are sophisticated and complex computer controlled robotic vehicles that are able to navigate their environment without the assistance of any form of guidance mechanism such as wire, reflectors, or tape. They are capable of examining their surroundings and avoiding people or obstacles. AMRs use sensors, artificial intelligence, machine learning, and software to plan and adjust their path when programmed to complete a task.
AMRs identify and map their surroundings and are able to “SEE” walls, fixtures, columns, and shelving. They use the collected data to navigate their environment using SLAM, which is a set of algorithms used to navigate by mapping and locating their position.
SLAM is a generic term that is used to describe a wide array of algorithms and technical approaches. The various types of SLAM include graph, EFK, fast, topological, visual, 2D and 3D LiDAR, and oriented fast and rotated brief (ORB) SLAM. The essential purpose of SLAM is to assist robotics to locate their position and orientation on a map as they create a map to complete their assigned tasks.
The concept of SLAM has existed for many years and can be seen in science fiction movies. In order for SLAM to become a reality, various technological advances had to occur to increase computer processing speed and lower the cost of sensors and scanners. The abilities of SLAM are dependent on two forms of technology, which are sensor signal processing and pose graph optimization.
PGO is a technique used to improve poses, which are positions and orientations, of a robot or camera. Its goal is to limit the number of errors of poses by considering their relationships and constraints. Pose graph is a graphical representation of the poses with nodes to represent poses and edges to represent spatial constraints that are provided by odometry and loop closures.
Odometry constraints estimate the motion between poses while loop closures are constraints in regard to a location an AMR revisits. By correcting and recognizing the loop, consistency is developed between the current and previous observations of a location. Pose graph optimization finds the best set of poses to satisfy constraints as close as possible. Optimization adjusts poses to reduce errors to ensure accurate and consistent poses.
The initial stage of implementing an AMR includes the use of SLAM to help an AMR map its surroundings. Using a joystick, an operator guides the AMR through its workspace. During the process, the AMR’s scanners detect the location of walls, equipment, machinery, and stationary elements in its environment that are saved by SLAM. This initial process includes a tour of every aspect of the workspace that the AMR will encounter. The efficiency of SLAM requires one tour of the workspace to have a complete map.
As changes and alterations occur in the workspace, the AMR has its SLAM map changed to accommodate new developments. Objects and equipment added to the workspace require that the mapping be dynamically changed to acquiesce to new guidelines.
During the mapping process, the AMR moves through the workspace taking snapshots of its surroundings. The picture that is created is formed by features represented as points in space, which include the distance from an object. The collection of feature points and their position are grouped together to form a cloud point, a 3D representation of the environment. The AMR tracks its position and orientation in the cloud points to enable localization.
Although mapping is a very impressive feature of SLAM, it is just the beginning of its use. Once the environment is mapped, SLAM helps an AMR understand its location from what it sees around it. Knowing its location, an AMR can create its path for completing its assigned functions.
When an AMR is activated, its first priority is to determine and identify its position and location. To complete this function, it uses cameras, sensors, and LiDAR to identify its place on its map. In some cases, GPS is used as further assistance. The localization process involves the interpretation of camera images that are delivered at 30 or more frames per second. With each frame, SLAM estimates distance, matches frames to previously tracked features, checks against the map, adds new features, and localizes the AMR’s position.
Visual SLAM uses cameras and sensors to help an AMR gain a visual interpretation of its environment. The types of visual SLAM can be as simple as a single camera or include compound eye cameras and RGB-D cameras. The various types of cameras are used for landmark identification and detection. Visual SLAM uses complex and complicated algorithms such as PTAM and ORB-SLAM, which are sparse methods for matching feature points of interest. Algorithms, such as DTAM, LSD-SLAM, DSO, and SVO, are dense methods that use the brightness of images for navigation.
LiDAR uses laser sensors, which are more precise and can be used with high-speed AMRs, such as self-driving cars. The laser sensors provide distance measurement and are very effective in map construction. Calculated movement or traveled distances are used to localize the position of the AMR. LiDAR uses point cloud matching with registration algorithms like iterative closest point (ICP) and normal distributions (NDT). The point cloud maps are represented as grid maps or voxel maps.
For LiDAR to be the most effective, it is fused with other measurement methods such as wheel odometry, global navigation satellite system, and IMU data. In workspaces where there are few obstacles or a great distance between obstacles, LiDAR may not be able to function without assistance.
The MiR600 has laser scanning technology that gives it 360o visual access to its environment for optimal safety. It is capable of picking up, transporting, and unloading pallets without the use of additional guidance systems. CAD files can be downloaded into it with maps of a facility, or MiR600 can create its own map. The MiR600 is an IP52 rated AMR with the ability to withstand the effects of dust particles and fluids and can drive close to fences and open gates where it may be exposed to droplets of water. It is controlled with an intuitive MiR Robot interface using a smartphone, tablet, or PC and can be programmed without prior experience.
The MiR250 hook is used for towing heavy products in manufacturing or moving food and linen carts in hospitals. The hook on a MiR250 hook is capable of supporting loads up to 500 kg (1100 lbs.) and provides a new and advanced logistics option. The MiR250 hook identifies carts by apriltags and moves the carts to a predefined location. The commands for MiR250 hook AMRs can be instantly changed using a smartphone or tablet and standard Wi-Fi to access the AMR’s control software. The strong robot base of the MiR250 hook makes it more maneuverable and assists in improving its performance.
The OPEX sure sort system is a scalable and cost effective method for handling multi line orders, package sorting, and reverse logistics. It is an ideal sortation system for handling small packages of any shape weighing up to five pounds. Sure sort requires minimal package touches and uses a six sided scan tunnel such that all barcodes can be read regardless of their position. It is the perfect system for small businesses that need a cost effective solution or large businesses looking to streamline their operation.
Sure sort is a put wall where an operator places items on a belt that moves the items through the six sided scan tunnel. Items are placed in an iBOT that is a multidirectional vehicle that deposits items in a bin location. When a bin has all of the items for an order, the operator is alerted that an order is ready to be packed for shipping.
The K55 pallet stacker is designed to move and stack palletized loads at a low height and can perform cyclical or conditional routes by interacting with other AMRs, systems and people. It is the modern automated solution for transporting and organizing medium weight palletized orders. The K55 pallet stacker is adaptable to any pallet storage application, merchandise reception, and material handling system. It optimizes storage space and improves process efficiency. The K55 pallet stacker can lift 1000 kg (2204 lbs.) to a height of one meter. It uses mapping software and has exceptionally high accuracy and precision. For safety, the K55 has 360o laser scanners with PLC safety and led signaling and front touch monitoring for AMR status, potential errors, and circuits.
The ODM is designed to carry totes or small loads up to 300 kg (661 lbs.) and is well suited for the electronics and pharmaceutical industries. It is able to navigate an environment without the need of modifying the workspace. The ODM uses an omnidirectional drive that allows it to move into rack aisles and turn, instantly. It has a route finding system that makes it possible to avoid obstacles or people. When a route is blocked or impassable, the ODM calculates a new efficient route to complete its task. A key feature to the ODM is its swarm application that allows a fleet of ODMs to communicate and share data.
The main focus of the AMR industry is to assist their customers by providing solutions that improve employee efficiency in regard to picking, locating, and moving products and inventory. AMRs are able to work continuously, which helps eliminate downtime and enhances productivity. Additionally, picking AMRs are exceptionally accurate, a factor that reduces customer returns.
AMRs improve operational efficiency and streamline workflow, which eliminates the need for manual intervention. They have advanced warehouse routing systems that reduce material handling and transportation that lowers energy use and an organization’s carbon footprint. AMRs work productively around the clock without any need for breaks or rest.
Each task that an AMR performs is precise, ensuring consistency and the removal of human error. Since they monitor production in real-time, they are able to identify bottlenecks, inefficiencies, and process errors, such that management can make corrections and respond to potential problems.
Inventory visibility is improved with automated inventory tracking and data collection tasks. Sophisticated sensors, cameras, and barcode scanners help AMRs to quickly and accurately perform inventory audits and real time updates of on hand stock and its location. The increased visibility helps maintain appropriate levels of inventory to avoid stock overages or stock outs and their costs.
The data collected by AMRs help businesses gain a vision of their inventory use making it possible to order appropriately and offer a vision of supply chain dynamics. By analyzing AMR data, it is possible to map patterns, identify trends, and prepare for business changes. New orders are processed with greater efficiency and in light of an item's usage. The end result is a more cost effective and productive operation.
A critical aspect of AMR use is their ability to alleviate employees of heavy duty tasks that can be risky and lead to injuries. This includes moving large bulky items that are physically demanding, time consuming, and potentially harmful. With AMRs handling heavy duty applications, the overall work environment becomes safer, which reduces injuries and frees staff to perform necessary functions that involve planning and problem solving. As AMRs take over pallet movement, employees can place more focus on quality control and order processing.
A side result of the use of AMRs is an improvement in the skill level of the labor force who can spend more time on enhancing their skills and developing practical solutions to everyday problems. Additionally, the use of AMRs has a positive effect on worker morale and attitude since they can take a role in strategic planning, invest in training, and develop programming for AMRs. As employees gain information regarding AMRs, they are more capable of adapting to innovations and changes.
One of the most time consuming tasks of an operation is order fulfillment and collecting. The use of AMRs radically improves the process and leads to faster and more accurate processing of orders. AMRs are used to pick, package, and ship orders efficiently while reducing the time needed to prepare an order.
It is very clear that the future of manufacturing, warehousing, and retail operations will be dependent on the use of autonomous mobile robots. Every organization, regardless of their business or size, will need to have a basic understanding of how AMRs work and how they can assist in improving normal operations. Over the last decade, business operations and methods have shifted, changed, and adapted to a wide range of technologies. This trend will increase as we move toward the middle of the century. Operations without AMR technology will be left behind.
Change is an inevitability and the nature of a successful business. Practices that were used a few years ago have become obsolete and replaced with more pragmatic and dynamic processes. The nature of AMRs makes it possible to adapt and change them to meet the ever shifting landscape of business. As a facility is updated and redesigned, AMRs are able to be programmed and reconditioned to meet the new dynamics.
An AGV forklift is a driverless self-operating robotic device that has the ability to carry, lift, retrieve, and place loads for easy transfer from one location to another. An automatic guided vehicle (AGV) forklift is a computer controlled mechanism that...
Automated guided vehicles (AGV) or mobile robots are types of guided robotic systems that are not bounded by a fixed range of motion. Rather, it is self-contained and can move along a line, surface, or space...
In 1954, when Arthur "Mac" Barrett, of Barrett Electronics Corporation, unveiled the first AGV, he named it Guide-o-Matic and described it as a driverless vehicle...
Palletizing is the process of putting items on a pallet. The process of emptying the loaded objects in the reverse pattern is known as depalletizing. A pallet is a flat, square-shaped platform used to transport and...
A pallet stacker is a machine designed to assist the user in lifting, moving and handling palletized materials with ease. A pallet itself is a flat and horizontal structure used to support goods in a sturdy fashion...
A palletizer is an automated material handling machine used to stack and orient several individual products into a single load for a more convenient and economical method of handling, storage, and shipment. Palletizers are usually part of a bigger packaging process...
A robotic palletizer is a type of palletizer that employs a robotic arm to pick, orient, and place individual products and arrange them into a single stack of load. They are the next generation of palletizers, and they will supersede conventional palletizers...