The autonomous navigation and collision avoidance capabilities of modern warehouse robots (AMRs) are achieved through their internal motors, software, and sensors.
First, AMRs are equipped with a motor system to drive and control the robot's movement. These motors generate power based on received instructions, enabling the robot to move and maneuver in a warehouse environment. The motor system typically consists of multiple motors used to control different parts of the robot, such as wheels or toes.
Secondly, AMRs rely on highly intelligent software systems to achieve autonomous navigation and collision avoidance. These software systems, based on advanced algorithms and planning techniques, generate real-time maps by fusing sensor data and determining the robot's position and obstacles in its surrounding environment. Based on this information, the software can plan optimal paths and actions to avoid collisions with workers, equipment, or other obstacles.
Finally, the AMR is equipped with a variety of sensors for sensing and perceiving its surroundings. These sensors can include laser sensors (such as LiDAR), visual sensors (such as cameras), and ultrasonic sensors. Through these different types of sensors, the AMR can acquire information about its surroundings, the location and distance of obstacles, and more. This sensor data is then processed and analyzed by the software system to achieve precise navigation and collision avoidance.
In summary, the autonomous navigation and collision avoidance capabilities of AMRs are achieved through the combined effects of a motor system providing power and control, a software system for path planning and motion control, and a sensor system for perceiving the surrounding environment and providing real-time data support. This combination enables AMRs to navigate autonomously in congested logistics centers, avoiding collisions with workers and equipment, thereby efficiently completing cargo delivery tasks.
Winning the game slowly and steadily
When you take a closer look at AMRs, the first thing you learn is that these machines aren't built for speed; they don't have acceleration from zero to 60. Instead, warehouse robots are designed to be slow and steady, moving inventory from point A to point B at a moderate pace while avoiding collisions with shelves, forklifts, and especially pedestrians. Safety is paramount, so most loaded AMRs—like Locus Robotics' Origin robots—cruise at a low speed of 2.5 miles per hour.
Acceleration and speed control
AMRs are designed for slow and steady movement, rather than high-speed acceleration. Their goal is to transport goods from point A to point B at a moderate speed, avoiding collisions with racks, forklifts, and pedestrians. This design prioritizes safety, so most loaded AMRs cruise at low speeds, such as 2.5 miles per hour (approximately 4 kilometers per hour).
Payload capacity
Different models of AMRs have different payload capacities. For example, Geek Plus's P40 model can carry up to 88 pounds (40 kilograms), while its P1200 series AMRs can handle loads of up to 2,645 pounds (1,200 kilograms).
Battery System
Most AMRs rely on one or two rechargeable lithium-ion batteries as their power source. These batteries are similar to those used in smartphones, but are heavier and have a higher voltage. Using such batteries is reasonable given that AMRs require a significant amount of energy to carry cargo, power sensors, and maintain up to eight hours of operation. When the battery is low, the AMR automatically navigates to a charging station for a quick charge, known as "opportunistic charging," or performs a full charge during extended rest periods such as overnight or between shifts.
Onboard computer
An AMR houses a critical computer. This computer not only monitors power levels but also communicates with the Warehouse Management System (WMS) or other software platforms. It retrieves the AMR's operating commands and reports back to the system upon completion of each task. These computers typically connect to the facility's wireless network via a Wi-Fi antenna on the AMR.
Vehicle-specific calculations
AMR performs other calculations directly on the vehicle itself, without relying on a wireless connection. These calculations are typically safety-critical calculations related to navigation or collision avoidance. This design aims to prevent system risks due to loss of wireless connectivity (such as a disconnection or power outage).
By combining these key components and technologies, AMRs can perform transportation and delivery tasks in a safe and stable manner in warehouse environments, while avoiding collisions and achieving effective communication and coordination with warehouse management systems.
AMR's sensing capabilities
When you delve into the inner workings of AMR, you'll be surprised to discover that its technology extends beyond batteries, computers, and antennas, encompassing an impressive array of sensors that essentially act as the machine's eyes and ears.
These sensors vary greatly in complexity. On simple machines like Automated Guided Vehicles (AGVs), sensors may be inexpensive devices that rely on external infrastructure for navigation. In other words, these sensors need some "landmarks" to orient themselves; these landmarks could be the positions of infrared beacons on shelves and walls, stripes on the floor, or even quick response (QR) codes to identify specific locations.
Kent Kjaer, a sales engineer at Danish AMR developer Mobile Industrial Robots ApS (MiR), explained that in more complex vehicles like AMRs, sensors are often more advanced devices capable of autonomous navigation. He said, "If an AMR detects a cart on the ground, it won't stop immediately; instead, it will go around it, or even reverse to choose an alternative path. Therefore, it requires a LiDAR scanner and a 3D camera."
LIDAR (Light Detection and Ranging) sensors can see a two-dimensional (2D) world at a height of about 8 inches above the ground. This is sufficient for a robot to determine its relationship to its surrounding physical environment, including walls, doors, and shelves—a process known as localization. However, this is not enough to provide a safe collision avoidance solution: the robot may still miss a pallet on the floor or a forklift fork outside its detection range.
To enhance the perception capabilities of AMRs, developers typically add 3D camera sensors that can map the environment around the robot within a range of approximately five to six feet above the ground. For example, a MiR combines a front-view 3D camera and a rear-view 3D camera, both with a 270-degree field of view (FOV), fusing their inputs together to form a complete 360-degree image.
In addition, MiR incorporates proximity sensors at various corners of the AMR to detect surrounding objects that the camera might miss, such as items just placed on the ground. As the AMR moves through its environment, it fuses the inputs from these multiple sensors into an image refreshed multiple times per second, using a process called Simultaneous Localization and Mapping (SLAM) to help the AMR avoid collisions and find appropriate paths. By integrating data from multiple sensors, the AMR gains comprehensive and accurate environmental awareness for safe and efficient navigation and operation.
Does each application have a sensor?
When designing an AMR, engineers can choose between mechanical lidar sensors, which measure time-of-flight (TOF) through laser emission and reflection from a rotating mirror, or solid-state lidar sensors, which have no moving parts. These sensors can be adjusted according to the robot's rolling speed to provide early warnings and allow sufficient time to stop, enabling them to "see" further distances and obstacles, even when carrying heavy loads.
In addition, there are lidar sensors specifically designed for outdoor use, enabling navigation in low-visibility conditions such as rain, snow, and fog. There are also sensors designed for robots working in cold storage areas and freezers, designed to prevent fogging.
SICK also offers a "pathological safety system" that collects data from a laser scanner and analyzes it using software on the AMR to adjust the driving speed, helping the robot avoid collisions and breakdowns.
As AMR applications develop, robot developers are constantly introducing new sensors and features to improve product performance. Some robots are now equipped with digital cameras to take pictures of barcodes and inventory, while others are equipped with "height sensors" to retrieve goods from high shelves. Still others are equipped with car-like headlights and horns, emitting warning sounds and flashing lights before turning to alert warehouse workers to their presence.
As logistics demands continue to evolve, AMR manufacturers will continue to improve technologies such as sensors, batteries, and computers to meet growing needs and create robust AMR systems that can adapt to the ever-changing logistics environment.
