Industrial robots are a special type of industrial equipment that can perform manual tasks efficiently, accurately, and repeatedly. They have been used in factories and commercial facilities since the 1960s. With the advent of Industry 4.0, these robots are integrating intelligence and new functions, making smart factories possible.
In addition to being more flexible in performing tasks, industrial robots can collect and analyze data about themselves to improve productivity, service quality, and reliability, while reducing total cost of ownership. When connected to the cloud, the operating patterns and trends of all robots can be identified.
For example, faults can be analyzed to create predictive maintenance algorithms that trigger alarms when there are errors in the robot's operation configuration log files. This allows problems to be resolved before they lead to equipment failure and downtime, thereby minimizing losses.
The growth of industrial robots continues as more and more industrial automation OEMs invest in robotics technology. Since 2016, the robotics industry has grown at an annual rate of 7.6% (CAGR).
1. OEMs continue to invest in industrial robot technology, and the robotics industry has grown at an annual rate of 7.6% (CAGR) since 2016, even taking COVID-19 into account.
In the coming years, industrial robot developers will face a range of challenges. As these systems become increasingly complex, reliability becomes paramount. This article will explore the challenges of collecting operational data from industrial robots in the form of data logs, including how to process the data that must be collected and how to minimize data loss during system failures.
The complexity of industrial robots
In its most basic form, an industrial robot consists of a robotic arm and a controller. The robotic arm, often referred to as a robotic arm, can move, rotate, and perform actions.
2. Block diagram of an industrial robot.
The various parts of each robotic arm are connected by mechanical joints, each joint providing a motion axis. A typical robotic arm has six movable joints, or six motion axes.
Each axis, controlled by a high-precision servo or stepper motor, is limited to a specific range of motion. Furthermore, each axis moves at a different speed, typically listed in degrees per second in a datasheet. The greater the range of motion, the faster the maximum speed of the joint, and the higher the precision required to control the motion. The need for greater coordination and precision also increases the amount of operational data that needs to be recorded from each sensor on the tracking robot.
From a reliability perspective, industrial robots must be able to recover from various power events, such as power outages. Ideally, once the power failure is resolved, the robot can immediately resume operation from its stopped position, even if the system has been reset.
Therefore, each motor must be able to save key parameters and data status, including the rotation angle and position of the robotic arm. Similarly, the controller needs to maintain detailed control logs, recording the operating parameters of each axis, including its command position, encoder value, and payload.
In addition, the controller must keep track of the servo motor's recorded tracking speed, torque, motor feedback sensing (i.e., current, position, speed), and motion angle. Reliably recording all this data requires some form of non-volatile memory so that the data is not lost due to power outages.
Non-volatile memory for data recording
For decades, critical data has been stored in battery-powered SRAM. However, this approach has many drawbacks:
It requires several components (battery, power management controller), takes up more PCB space, and increases the number of potential points of failure.
To prevent the battery from overheating, it is usually necessary to install the battery after the reflow process, which increases manufacturing costs.
Industrial robots are often exposed to vibration, which can cause mechanical failures in the connectors that hold the battery in place, thus reducing overall reliability.
The battery needs maintenance and replacement.
The batteries do not meet RoHS requirements, which has created disposal issues for the manufacturer.
For these and other reasons, OEMs have turned to non-volatile memory devices to replace battery-powered SRAM. This table shows a range of non-volatile memory technologies available to OEMs.
Due to the low durability of EEPROMs, they can be excluded from most applications. Industrial robots operate 24/7 and must record large amounts of real-time data. Since these robots may run continuously for years, EEPROMs will eventually wear out, making them an unfeasible option.
Flash memory also has limited battery life. However, the durability issue of flash memory can usually be addressed using wear-leveling software techniques on the host processor. When a block begins to experience errors exceeding a set threshold, the wear-leveling algorithm moves the data to a more reliable block.
Wear leveling effectively extends the lifespan of memory by distributing wear evenly throughout the flash memory. However, the process of tracking and moving data throughout memory increases the host CPU load and introduces write operation latency.
When using flash memory for data logging, the most important consideration is likely that it writes data in blocks. Log data must be collected in a buffer until the entire block is ready to be written. Wear leveling algorithms may involve software-based lookups in a large table to select the blocks where data should be written. Finally, the flash memory must erase the blocks before writing.
Only after these tasks are completed can log data be finally written. All of these factors contribute to a significant delay between actual data acquisition and writing.
Real-time reliability
As mentioned earlier, the two main reasons for data logging are performance analysis over time and power event recovery. For both of these functions, arguably the most important information is the data collected during a failure.
In the event of a power failure, the data will be used to recover and pinpoint the exact location where the industrial robot stopped operating. For performance analysis, this "last-minute" data is crucial for understanding what happened before the failure and what possible causes might have led to it.
When a system malfunctions or experiences a power failure, there is almost no time to react. With flash memory and EEPROM, anything in the buffer will be lost. However, this is the most critical data. The longer it takes to write to memory, the greater the risk of losing critical data. Consider a high-precision robot operating on expensive components. If the robot encounters a power failure, the system needs to be able to reset with high precision to the interrupted position. Otherwise, the part being processed may be scrapped.
To maintain highly reliable operating parameters and data logs, data must be continuously captured and stored in non-volatile memory. For this reason, robot developers are turning to ferroelectric random access memory (F-RAM). As can be seen from the table, F-RAM offers many advantages, making it the preferred choice for storing critical operating parameters and data records.
F-RAM boasts an endurance of 10 to the power of 14 write cycles, providing virtually unlimited durability for data logging applications. Furthermore, F-RAM does not require wear leveling, thus simplifying and reducing write latency.
F-RAM is a type of random access memory that does not require refresh cycles. No data blocks need to be buffered because data can be stored immediately in non-volatile memory. Furthermore, F-RAM's random access characteristics eliminate the latency associated with memory paging. Data is stored immediately when it is captured.
Data records market trends
Developers must decide whether to centrally record data within the main controller or at the edge of each motor. Currently, data recording at the motor edge requires up to 1 Mb of capacity, while the controller requires up to 16 Mb.
For high-speed applications such as six-axis robot controllers, Infineon's latest generation of non-volatile memory, Excelon F-RAM, offers higher-density memory and a four-channel SPI interface to help improve throughput. For applications with lower data logging requirements, there are lower-density single-channel SPI products available.
However, as the number of axes and sensors in industrial robots continues to grow, the requirements for data logging will only expand (Figure 3). At the same time, AI-based performance and predictive maintenance algorithms will require access to a wider range of parameters with greater granularity, thus increasing the total amount of data that must be collected and stored.
3. As the number of axes and sensors in industrial robots continues to increase, the requirements for data logging will grow over time.
Another trend influencing non-volatile memory density is moving recording functions closer to the network edge. Highly reliable and functionally safe edge computing and storage in each motor can eliminate latency in sending data back to the main controller.
Many manufacturers use microcontrollers on each motor, whose movements are coordinated by a master six-axis controller. Thus, each motor tracks its own parameters and sensors. This, in turn, enables more advanced artificial intelligence and machine learning (ML) capabilities to be transferred to the edge and to individual motors.
Other storage devices in industrial robots
In addition to data recording memory, industrial robots also employ many other memory technologies in their systems, including storing boot code as extended memory. With the advent of Industry 4.0, the need to protect systems from cyber threats has surged.
One of the primary targets of hackers is flash memory devices, which store boot code, security keys, and other critical data essential for the normal functioning of the system. In response, Infineon has developed SEMPER Secure NOR flash memory, which conforms to functional safety standards and integrates security features to protect the code from hacker attacks.
The increasing complexity of robot controllers has led many to also incorporate their own TFT displays to support direct interaction with technicians and remote control. HyperRAM is well-suited as extended memory for industrial displays, serving as buffer for data, audio, images, and video, or as temporary storage for mathematical and data-intensive computations. It offers transfer rates up to 800 MB/s on a low-pin-count serial HyperBus interface.
in conclusion
Data logging is a fundamental function of industrial robots, enabling recovery from malfunctions and power events without negatively impacting production. It also plays a crucial role in enabling emerging AI and ML capabilities, such as predictive maintenance, by providing data that will drive innovation in these applications.
The infinite endurance of F-RAM, combined with its real-time performance, non-volatility, high throughput, and reliable data capture, makes it a powerful choice for non-volatile memory for high-performance data recording in industrial robots. Because F-RAM guarantees minimal data loss during power outages, it enables high-precision recovery, allowing robots to continue operating from where they stopped before a reset or failure.
F-RAM is available in low-density and high-density options to meet the requirements of different applications. This also gives developers the flexibility to meet the evolving needs of next-generation robots as AI and ML capabilities move closer to the edge.
