The supply chain as a whole has never been more of a focus than in recent years, with disruptions wrought by the pandemic impacting consumers and businesses in nearly every industry across the globe. This has caused more companies than ever before to accelerate investments in automation technology. When evaluating new automation solutions, buyers quickly realize not all automation technologies are created equal. In the case of autonomous mobile robots (AMRs), today’s AMRs operate differently and vary in their degrees of autonomy as well as their reliability, flexibility, and functionality. Choosing the right mobile automation solution can help create a resilient, agile, and safe operation that can help warehouses and manufacturing facilities adapt in today’s ever-changing environments.
Traditional vs. Nontraditional Automated Solutions
There are different categories of self-driving industrial vehicles, including autonomous mobile robots (AMRs) and automated guided vehicles (AGVs). These material handling solutions serve a similar purpose of moving materials within manufacturing, warehousing, or logistics facilities, but how they operate is where they differ. Traditional AGVs follow fixed paths that are defined with physical infrastructure such as tape, wires, magnets, or reflectors. The downside of infrastructure-based automation solutions is that they can hinder a facility's ability to remain flexible and adopt a continuous improvement mindset. The friction of change, costs associated with reconfiguring a facility, and risk of downtime when making updates can be a competitive disadvantage. The continuous need for supply chain agility has led many companies to turn to highly flexible AMR solutions.
AMRs are similar to AGVs in that they are also driverless industrial vehicles; however, AMRs can navigate entirely through onboard sensors and technology, meaning they do not need any external landmarks to move and operate within a facility. This advantage means that AMRs sense their surroundings in real-time, navigating based on the data they gather to drive reliably and reduce downtime. It’s also important to look at the different sensors and technologies AMRs rely on to gather that data, and whether they are constrained by their sensory abilities or by the software that makes decisions on how to use the data.
Many AMRs today rely on Light Detection and Ranging (LiDAR) sensors to navigate. This method uses a laser-based sensor to help provide AMRs with a sense of spatial awareness. The sensor takes direct measurements of the environment around the AMR by sending out pulses of laser light. It then times how long it takes for the light to return to the sensor to calculate the exact distance to the object detected. By scanning an area, it helps create a geometric map of the space that the AMR navigates based on the location and orientation of the objects around it.
The LiDAR sensors used in AMRs have two different categories: 2D and 3D. As implied, 2D sensors operate in two dimensions and use a single plane of lasers to capture data and have a very limited view of the surroundings. Similarly, 3D sensors capture data across X and Y dimensions, but these sensors also can take additional measurements across the Z axis to collect three-dimensional data. Individual measurements through LiDAR are very precise, however they have limited vertical resolution, which leaves gaps at longer distances and can challenge an AMRs’ ability to perform in highly dynamic environments.
Computer Vision Navigation
Computer vision is an interdisciplinary scientific field which seeks to understand and automate tasks that the human visual system can do. A camera can be used as a sensory input when combined with software, and can be programmed to help a mobile robot “see”, analyze, and comprehend the content in its visual world. The cameras used in computer vision are able to collect robust amounts of environmental data because of their wide field of view, high resolution, and ability to distinguish textures, visualize colors, and recognize objects.
Configuring the cameras in pairs gives vision-based AMRs a significant advantage when navigating because it provides depth perception in a way similar to how humans view the world. Using multiple pairs of cameras mounted around the AMR enables stereoscopic vision, helping it see a three-dimensional environment, but with a more expansive, even hemispheric field of view. Individual measurements are less precise, but resolution and coverage create very dense and consistent data, which when captured and processed in real-time creates high reliability in dynamic environments like industrial settings.
Leveraging the Combination: Computer Vision and LiDAR
The best way to take advantage of vision (cameras) and LiDAR is to combine them for a hybrid sensor approach. This leverages the strengths of both technologies to create not only the most reliable AMR, but also the safest. Fusing data from computer vision cameras with LiDAR sensor data delivers an unmatched level of AMR performance, collecting a higher density of information that the software can prioritize and filter. The advanced software enhances the AMR’s understanding of its immediate surroundings by interpreting data points. The more detailed and expansive the AMR data is, the better its ability to comprehend and make decisions across a greater variety of situations and sort through the inbound information the robot receives. Robust amounts of information ensures reliability as opposed to trying to extrapolate from a limited set of data points.
Any disruption in the supply chain workflow can result in major losses in downtime, productivity, profits, and overall competitive advantage. To tackle these challenges, investing in automation is key, but choosing the right technologies to invest in can make or break operational goals. Ensuring the automation investment is equipped to collect the data needed to make time-saving decisions is key. These technologies will empower and encourage facilities to deliver continual improvements while providing a return on investment in the form of efficient and repeatable movement, improved safety, and reduced downtime. In turn, manufacturers will be able to achieve their operational goals today and remain flexible in the future.