What is a Mobile Robot?
Let’s dig into the details of exactly what a mobile robot (or AMR) exactly is. We’re going to look at the basic details that define an autonomous mobile robot. Along the way, you’ll learn the basic things that will help you differentiate the various vendors and solutions available on the market. Armed with this information you can make a more informed buying decision.
Mobile robots are a unique combination of mechanical and electrical systems, together with innovative software capabilities. The basic designs of a mobile robot system are no longer rocket science. The magic now resides in the software applications which supervise a mobile robot fleet and which enable the mobile robots to perceive their world and navigate effectively and efficiently.
As proof of this, around the world, hundreds of thousands of high school students compete annually in the FIRST robotics competition. If high school students are capable of taking a $500 box of parts and building a machine which can maneuver around a playing field, you can quickly see why the basic mechanical design of mobile robots are no longer a differentiator.
In this section, you will learn about:
- The types of mobile robots
- Carrying capacity and payload design considerations
- Safety Requirements for mobile robots
Types of Mobile Robot Applications
The simplest way to classify autonomous mobile robot applications is to compare them to the way that human driven vehicles operate on a road. Most mobile robots operate in some form of the following traffic patterns.
Long haul, bulk material movement
Queuing and consolidating material for movement is a core concept that has been employees for centuries. From the camel caravans of the silk road to modern day freight trains and merchant shipping, there are many examples to consider.
Long haul trucks and railroads are designed to carry large, consolidated loads over long distances. A railway system stays on a predetermined track and always includes multiple “cargo cars” which are optimized for different types of loads. Long haul material movement typically operates on an “as-needed” schedule.
In a similar fashion, human operated tugs on the factory floor or warehouse may move a train of carts throughout the facility to pick up and deliver material as it’s needed at various locations within the facility. Autonomous mobile robot “tuggers” are one option available today to solve this requirement with a facility.
Scheduled Routing
A bus system is designed to operate along a fixed route, stopping at predetermined ‘bus stops’ on a predetermined schedule. Buses often have their own special lanes to separate them from other traffic and enable them to keep to their predetermined schedules. Passenger trains and commercial airliners also operate with this design, over longer distances.
Automated guided vehicles (AGV) have been in operation on the factory floor since the 1950’s. Classic AGV’s follow a similar format of travel in the facility. By following a magnetic tape track on the floor (similar to a railroad track), the AGV’s always transit the facility in the same direction, constrained to the paths defined for them.
Mobile robots can also be assigned to patrol a specific path on a schedule or to move between specific “bus stops” along a route. Mobile robots can perform this operation while using the same walkways and aisles used by humans.
A variation of this might be called “local delivery”, similar to the neighborhood delivery of packages by the postal service or package delivery services. In this example, the vehicle carries a load of packages to a neighborhood and stops at individual locations to drop off specific material.
On-demand Taxi Service
Taxi’s, on the other hand, are designed to operate with an “on demand” schedule. Picking up passengers from any point and dropping them off at any other point. Taxi’s may congregate at well known pick up locations such as a convention center or airport while waiting for a fare.
Like these classic transportation schemes, modern mobile robots have been designed to operate in similar fashion. As a result, modern mobile robots require additional intelligence and several types of sensors to navigate the environment autonomously. Mobile robots also require an accurate map of the facility so that they can determine the best path from one point to another.
Summary
The type of delivery path will determine which type of automated vehicle is the best fit for your application. If the path requires randomness in the execution, then an autonomous mobile is likely the best fit. However if the path can be constrained to a well defined path, then an automated guided vehicle is likely a good candidate.
Automated Guided Vehicles
Automated Guided Vehicles (AGV’s) have been around since the 1950’s. This class of vehicle was designed to operate like a bus, running around a facility in a predefined path (or lane), and (typically) on a predefined schedule. AGV’s have been designed to carry heavy payloads such as automotive chassis’s, airplane components, engines or other large manufactured parts throughout the facility, and they still have applications on the modern manufacturing floor.
There are several mobile robot vendors in the market today who erroneously refer to mobile robots as AGV’s, and there are several AGV vendors who try to position their solutions as mobile robots. We’re not going to perpetuate this classification. AGV’s are as distinctly different from mobile robots as a bus is from a taxi.
“A mobile robot is NOT an automated guided vehicle”
AGV’s use state of the art navigation technologies such as magnetic tape, magnets embedded in the floor, or laser targets/fiducials to navigate and drive the vehicles safely around a facility. Most AGV’s are capable of sensing an obstacle in their path and avoiding hitting it, but they are not capable of moving around an obstacle blocking their path. An AGV will typically raise an alarm and enter an emergency stop situation which faced with an obstacle in its path. As a result most facilities set aside specific AGV paths throughout the facility, and expect that humans, fork truck drivers and all other aisle track keep out of the AGV pathways. This is similar to any conveyor system, which will occupy a specific space in the factory or warehouse facility. Most AGV systems are very robust in their design and operation as they have had many years to evolve and harden the AGV designs.
AGV can also be deployed to move “carts” around a facility. In this use case, the AGV, often described as a “Tunneling AGV”, moves under a wheeled cart, connects to the cart and transports it to another drop off point where the material can be unloaded from the cart. The tunneling AGV can drop one cart off at the end of a queue, and then drive under the queue of cart and hook up with the cart at the end of the queue.
AGV systems are generally less expensive than a mobile robot of comparable size and operating payload. This is due in part to the lower complexity of controls (I.e. control system) and simplicity of design of an AGV compared to a mobile robot. AGV’s operate under battery power, just like a mobile robot, however most AGV’s use mature battery chemistry which is less expensive than that of modern mobile robots. In addition, AGV’s which operate continuously will have automatic battery swap stations so that one set of batteries can be charging while another set is running the vehicle. If an automated battery swap is not a feature, then the AGV must be managed by human operators who will swap battery sets as needed or take the vehicle offline to sit and recharge the batteries.
Why AGV’s are not mobile robots
AGV’s are not considered mobile robots due in part to their constrained navigation capabilities, limited sensor technology and lack of control intelligence. Likewise, mobile robots are not considered AGV’s since they are able to “roam freely” throughout the facility, access the environment condition, make independent navigation choices and avoiding obstacles with their on-board control systems.
Summary
AGV’s are useful when you have the floor space to allocate exclusively for the operation of the vehicle and when the application involves the movement of a large mass (like a car body or airplane wing) or when the process is structured to include the linear movement of material similar to a freight train.
What’s in a Name?
It seems like each mobile robot vendor has created a unique three or four letter acronym to define mobile robots and differentiate themselves from the rest of the market. Don’t be confused, all of these terms mostly mean the same thing. Here’s a sample of some of the definitions used by mobile robot vendors over the last 10 years:
- Autonomous Intelligent Vehicle (AIV)
- Self Driving Vehicle (SDV)
- Mobile Robot (MR)
- Motile Robot
- Smart Autonomous Mobile Robot (SAMR)
- Autonomous Mobile Robot (AMR)
- Autonomous Robot (AR)
- Autonomous Data Machines (ADM)
- Autonomous Delivery Robot (ADV)
- Robotic Material Handling Solution (RMHS)
- Unmanned Ground Vehicles (UGV)
- Vision Guided Vehicles (VGV)
The “A” has also taken one of several forms:
- Autonomous
- Automated
- Automatic
Autonomous Mobile Robot
To limit confusion, for the remainder of this guide we are going to use the most commonly accepted terminology with the term: Autonomous Mobile Robot (AMR) or simply Mobile Robot. Within some of the vendor product reviews, we may use the actual vendor term if it’s a part of their branding language, however note that within this report, these terms all refer to the same class of machines.
Finally, none of this should be confused with the emerging Autonomous Vehicle (AV) technology which has hit all of the headlines recently with the growth of self driving cars (and trucks). Autonomous Vehicles are automobiles or trucks which operate autonomously, or support the human driver in transporting humans and other cargo over public streets. This report does not cover the AV market at all.
However, the AMR market is set to profit from the advances in technology that the autonomous vehicle advancements will bring. AMR’s and AV’s both use similar sensor technology. AMR’s operate in more structured environments indoors while AV’s have to operate in highly unstructured environments at much faster speeds. In the end, the sensors and algorithms developed by AV’s will reduce the costs of AMR’s and improve their capabilities over time.