Autonomous Mobile Robots (AMRs) are gaining tremendous traction in the enterprise world. Whether a retailer wants to accelerate its last-mile delivery services or a micro-fulfillment center is dealing with staffing shortages, an autonomous robot is a viable option. While much of the attention of robotics, and automation in general, has been concentrated on warehouses and other indoor settings, AMRs excel in outdoor work environments as well. From energy companies to public safety officials, AMRs are expected to be an integral part of enterprise/organizational operations on a daily basis. Given these factors, ABI Research forecasts that the number of outdoor mobile robot shipments will grow at a Compound Annual Growth Rate (CAGR) of 27%, from 40,000 in 2021 to 344,000 in 2030.
Chart 1 below breaks down outdoor mobile robot shipments by market vertical through 2030.
Below is the YouTube video version of this blog post.
What Makes a Mobile Robot “Autonomous”?
To be considered autonomous, a mobile robot must meet the criteria for three key categories: navigation, manipulation, and functional safety.
Navigation refers to the ability of an AMR to move within its environment (e.g., warehouse, port, field, etc.) and to perform specific tasks like perception, awareness of the surrounding environment, localization, mapping, kinematics, generating the required insights, and mobilize the robot toward the target destination. AMR navigation can be done through three different approaches:
- Infrastructure Heavy: Two-Dimensional (2D) codes on the ground, camera-based sensing, Quick Response (QR) codes, etc.
- Infrastructure Lite: External camera system for remote control of robot movement
- Infrastructure Free: Relies solely on internal vision and laser sensors
Manipulation alludes to the AMR being able to execute physical tasks on objects in the operating environment, such as lifting, grasping, tilting, posting, or twisting. These actions are achieved via lifting platforms and industrial robotics, which rely on mechanical, electrical, and electronic components programmed for specific tasks.
Functional Safety (FuSa) involves software and hardware that enables the safe use of autonomous mobile robots. To facilitate the safe operation of AMRs, three factors are paramount: industrial robot safety standards, regulations and legislation that generate robot safety requirements in workplace settings, and the necessary technology.
Next, we will delve into some of the outdoor applications and use cases of autonomous mobile robots in various industries.
Oil & Gas
For mobile robot developers, oil & gas platforms are a prime opportunity for product innovation, given the challenging operating environments. Oil & gas platforms typically have mesh-wired stairs and tight passages, hazardous areas, cables and pipes, and wet or slippery surfaces. Wheeled AMRs would struggle in these work settings, which makes quadrupeds the optimal solution.
A quadruped’s design enables it to excel in tricky working environments. The robot’s hip and knee actuators provide significant degrees of freedom, allowing for improved mobility in tight spots. Moreover, a quadruped can duck under an obstacle or sidestep to avoid a potential collision with an object. Quadrupeds usually have five or more cameras, providing the robotic machines with more situational awareness, as well as temperature tracking and gas leakage detection capabilities.
Boston Dynamics’ robot Spot is perhaps the best example of a quadruped being used in industrial environments. Oil & gas juggernauts NP and Woodside leverage Spot to inspect areas where high-voltage transformers are located. In the event of an emergency, Spot can be the guinea pig, and go in first to collect visual cues that will help human response teams prepare. Alternatively, Spot is useful for site inspection during periodic shutdowns. Similarly, Malaysia’s national oil & gas provider PETRONAS deploys the ANYmal robot to inspect environments prone to explosions.
Figure 1: ANYmal performing an automated routine inspection of rotating equipment
Autonomous tractors automate agriculture tasks using cameras and Global Positioning System (GPS) (for pathfinding and localization). A mobile robot platform can automate human farmer operations, such as harvesting, mowing, weeding, seeding, and spraying. A vision-based Simultaneous Location and Mapping (SLAM) solution would not be ideal, as most farms have plants spaced uniformly, making GPS critical for AMR applications in these environments.
John Deere’s 8R tractor, launched in early 2022, is a prime example of an autonomous tractor. The front and rear cameras on the 8R tractor are embedded with computer vision (Machine Learning (ML)), which helps decipher between crops and weeds. This AMR also leverages Blue River Technology’s See & Spray machine, which processes images of plants at a rate of over 20X per second. Once the machine captures the images of the plants, they are cross-referenced across a training library of more than 1 million pictures. This technique is similar to how facial biometric authentication works.
If GPS is unavailable at an agricultural site, farmers can use a solution like Verdant Robotics’ NavBox, a platform with sub-centimeter location awareness. Instead of navigation, the company uses it for data collection and annotation, utilizing a computer vision system with Light Detection and Ranging (LiDAR) and a stereo camera housed in a dome-like casing. The system generates images for high-precision spraying, thinning, and weeding, which subsequently provides a season-long snapshot of fertilizer or weed control performance.
Construction and Mining
Given the extreme weather, temperature, and high level of pollutants/dust at construction and extraction sites, AMRs in these applications must be water-resistant, robust, and rugged. All autonomous construction and mining robots are built upon existing solutions, with additional hardware and software, such as proximity radar, 360° cameras, GPS, kinematic software, etc. These autonomous machines are powered by liquid-cooled computers and stored in an all-weather, ruggedized enclosure. Quadrupeds and drones often come in handy as well in construction and extraction sites.
Having deployed more than 250 autonomous trucks worldwide, Komatsu is a leading vendor in the construction and mining segment. Chinese startups, such as WAYTOUS, Tage I-Driver Technology, and EQ, are vigorously testing autonomous truck products. The Chinese government is a catalyst in these developments as it provides tremendous support for AMR projects. However, none of these trucks are fully autonomous, relying upon remote monitoring and teleoperation for mining trucks. In fact, completely autonomous mining trucks are unlikely to hit the global market for some time.
Finally, autonomous mobile robots are also being used by enterprises to automate the layout of a construction site using a Building Information Modeling (BIM) file. Robots will map out the data on the floor of a project site by simply referencing the BIM file and leveraging geospatial technology. This automated approach to building design facilitates improved collaboration among structural engineers, architects, Mechanical, Electrical, and Plumbing (MEP) engineers, designers, project managers, and contractors.
Last-mile delivery is another industry where AMR applications are gaining momentum, as supply chains contend with driver shortages, reduced delivery times, fulfillment complexity, and booming demand for e-commerce products. While some mobile robotic systems, such as those developed by Coco, are not autonomous, other vendors like Neolix, Nuro, and Ottonomy are developing fully autonomous delivery robots.
Like autonomous vehicles, these last-mile delivery robots are supported by a variety of sensors, including stereo, Time of Flight (ToF), radar, and LiDAR. While other AMRs are targeting the last-mile delivery industry that are semi-autonomous (Starship Technologies and Kiwibot), these systems are gradually headed for full autonomy.
Examples of last-mile delivery robots include Alibaba’s fleet that carries parcels from e-commerce sites, Amazon Robotics’ electric sidewalk delivery robots (which have been scrapped), and Coco’s remotely controlled sidewalk robots that travel at a speed of 5 Miles per Hour (mph).
Figure 2: Coco Sidewalk Delivery Robot
(Source: Coco Delivery)
As pointed out in ABI Research’s Seaport Digital Transformation report (AN-5550), robotics is a huge aspect of port automation. Whether it’s an unmanned transporter transferring containers or an Automated Guided Vehicle (AGV) transporting goods without a driver, automation is a key trend in the supply chain industry.
The Asia-Pacific region is a hotbed for mobile robots in ports. For example, the 130-vehicle fleet of AGVs at the Yangshan Port in China allows for unmanned driving, automatic navigation, path optimization, self-power monitoring, and proactive fault prevention. Down at the Port of Singapore, logistics operators are leveraging Scania and Toyota-supported autonomous truck platooning systems to move containers between terminals. Cloud-based Vehicle-to-Vehicle (V2V) communications results in a close arrangement of trucks, which leads to shorter time gaps and reduced latency between handoffs.
Konecranes is a big player in the port segment for AMRs. A market-leading Finnish crane and lifting equipment vendor, Konecranes has developed unmanned, automated container transport vehicles for rapid and economical container transport between the quayside and the container yard. The company also provides software for the navigation of autonomous vehicles.
Public safety is a highly complex and connected subject area, which requires everything from specialized monitoring technologies to personnel trained to respond accordingly. AMRs help automate public safety processes by capturing undesirable behaviors on city streets before alerting a command-and-control center. From there, human operators can make the appropriate judgment call. These autonomous wheeled robots rely on the following technologies and solutions to detect questionable behaviors in public:
- 360° High-Definition (HD) night and day video capture positioned at eye level
- Live streaming and recorded HD video capabilities
- Automatic license plate recognition
- Parking meter feature
- Vehicle dwell time tracker
- Facial recognition
- Thermal imaging
- Two-way communication
- Signal detection for sensitive areas
How Robotics Companies Can Capitalize on the Demand for Automation
When it comes to developing mobile robot platforms, there is a long list of companies active in outdoor applications. These vendors tend to target specific industries, as demonstrated in the blog post.
However, robotics companies must carefully consider several factors. For example, how does the vendor plan to implement 5G into the AMR solution? Or which type of AMR is best suited for a specific work setting? These market considerations, as well as many more, are discussed in fuller detail in the Outdoor Mobile Robot Platforms and Applications research highlight. This content is part of ABI Research’s Industrial, Collaborative & Commercial Robotics Research Service.