Automated guided vehicles have been used for many years to automatically pick up boxes and pallets, and carry them around factories and warehouses. While AGVs follow fixed optical or electromagnetic lines on the factory floor, autonomous mobile robots, AMRs, are now being used in a wider range of environments than the traditional factory and warehouses. AMRs are typically smaller and more agile to facilitate this type of operation.
Traditional applications for AGVs are obvious choices for the deployment of AMRs. These range from moving small packages around human sorters to moving heavy machinery and structures in large manufacturing. Changing trends in retail logistics are driving much of this innovation, with online sales leading to very complex inventories, product shipping direct from warehouses and processing of both orders and returns.
The improve flexibility of AMRs is also opening up new applications. In hospitals, AMRs are providing secure deliveries of pharmacy medications and laboratory specimens, as well as delivering meals, linens and environmental services. Hospital robots often look like a rolling cabinet, with drawers that the healthcare practitioners can access. Aldo Zini, CEO of Aethon, described the AMR it developed for hospital applications: “It’s completely autonomous. It can travel down hallways, communicate with its environment, ride elevators, open doors. It’s a very efficient, affordable and reliable technology that we’ve developed for the internal logistics in facilities, particularly in hospitals.”
Delivery robots have been trialled in several locations, including AMRs such as Starship in Milton Keynes. These robots can autonomously navigate pavements, avoid pedestrians and obstacles and cross roads, to make deliveries over several miles. Robo Mart is a larger car sized AMR that functions as a mobile fruit and veg shop. The hope is that these retail AMRs may enable small local businesses to compete with the convenience of major ecommerce businesses such as Amazon.
Amazon has trialled its own delivery AMR, known as the Scout, although it found this method of delivery didn’t work for its customers. FedEx also trialled its own AMR for deliveries, known as Roxo. Despite trialling the Roxo in the US, UAE and Japan, FedEx did not continue with this project. “Although robotics and automation are key pillars of our innovation strategy, Roxo did not meet necessary near-term value requirements.” Sriram Krishnasam, Chief Transformation Officer, FedEx.
In agriculture, AMRs can enable large tractors to be replaced by many small robots for many applications. This can reduce costs and environmental impacts by enabling greater precision. For example, rather than a tractor spaying weed control or pesticide indiscriminately over a field, AMRs can deliver the chemicals only where they are required. AMRs can also be used for harvesting, planting, pruning and thinning. For heavy tasks such as ploughing, full size tractors can also be made autonomous. In practice, general purpose field robots are not being widely used, but specialty robots are.
SENSORS AND ALGORITHMS
Although AMRs may be more physically agile than traditional AGVs, the real difference lies in the sensors and algorithms that enable autonomy. The sensors AMRs use to perceive their environment include digital cameras, radar and sonar. They typically use more than one type of sensor to provide accurate and reliable sensing in different environments and at different ranges. The most common sensors are listed.
Digital cameras are used for computer vision, with object recognition enabling products, people and landmarks to be identified Lidar (light detection and ranging) uses lasers to create a 3D model of the robot’s surroundings. These sensors can detect shape, distance, and location of objects. They are commonly used for obstacle detection and mapping. Radar works in the same way as Lidar but uses radio waves rather than lasers. This is typically used at longer range and for outdoor environments. Ultrasound sensors use high frequency sound to detect the distance and location of objects from their echoes. This is mostly used for obstacle detection and collision avoidance. Inertial measurement units (IMUs) combine accelerometers and gyroscopes to measure a robot’s motion and subsequently position by a process known as dead reconning. Such measurements are accurate for short periods of time but suffer from significant drift over time. They are therefore useful to fill the gaps between availability of other positional data such as landmarks.Robots use a hierarchy of algorithms to achieve autonomy, with higher level algorithms building on lower-level algorithms to provide an incrementally more sophisticated understanding of the environment and required actions. These typically progress as follows:
1. Feature recognition extracts lines, edges and shapes from 2D images, typically acquired by digital cameras.
2. Object recognition enables the robot to detect and identify objects in its environment. Object recognition is important for tasks such as material handling and inspection, as well as navigation and collision avoidance.
3. Simultaneous Localization and Mapping (SLAM) creates a map of the environment and allows the robot to localise itself within that map. SLAM is a key technology that enables AMRs to navigate autonomously in complex, dynamic environments.
4. Path planning algorithms enable the robot to plan a trajectory from its current location to a desired destination. They take into account factors such as obstacles, terrain, and other constraints to find a safe and efficient path.
5. Control algorithms enable the robot to maintain stability and perform tasks with precision. They can be used for tasks such as grasping, manipulation, and locomotion.
6. Machine learning algorithms enable the robot to learn from its environment and improve its performance over time. This typically involves reinforcement learning – a trial and error process involving rewards for successful actions and penalties for unsuccessful actions.
BENEFITS FOR SMALL WAREHOUSES
AMRs can work with humans or as teams of AMRs. For small warehouse operators, one of the biggest benefits is safety. Forklifts remain a major workplace hazard, causing 1,300 debilitating and life-changing injuries in the UK every year. Most of these accidents are caused by the human errors of drivers and could be avoided if AMRs were used.
This is of course in addition to the usual benefits of automation – AMRs can operate 24h a day without breaks, holidays or illness. They typically achieve faster cycle times than humans, and they are more accurate. Using barcodes to track and locate inventory, AMRs avoid the risk of logistical errors.