Powering Future Robotics via Programmable Logic Devices

Author : Karl Wachswender, Lattice Semiconductor

05 April 2024

Industry 4.0 is the impetus behind the rapid integration of autonomous technologies into manufacturing and industrial processing activities. Approximately 553,000 industrial robots were installed in facilities around the globe during 2023, according to the International Federation of Robotics (IFR) - and their prevalence will increase still further over the course of the next 12 months.

Worldwide spending on Industry 4.0 related applications - ranging from artificial intelligence (AI) and machine learning (ML) through to IIoT and cloud computing - is projected to increase significantly in the coming years. According to forecasts made by analysts from MarketsandMarkets, by 2026 it will represent $165.5 billion worth of annual revenue. This means a 20.6% compound annual growth rate (CAGR) will be seen over that period. 

As businesses invest more in Industry 4.0, there will be much greater reliance on robotics - thus enabling manufacturing throughputs to be boosted substantially and a simultaneous lowering of operational costs. However, there are factors that make developing next generation robotics more complicated. For instance, modern industrial systems now include greater numbers of sensors, multi-axis motors and cameras than ever before. They also require more versatile and secure microelectronic solutions in order to support enhanced cloud-to-edge connectivity.
 
As manufacturers look to augment their processes, flexible solutions will be needed to help align robotics automation with new operational expectations. Field programmable gate arrays (FPGAs) are clearly going to be pivotal in the design and implementation of Industry 4.0 hardware - enabling innovative and energy-efficient robots which are capable of attaining elevated levels of performance to be deployed on factory floors and in warehouse sites.
 
The state of robotics amid Industry 4.0
The robotics landscape is currently evolving, due to a massive influx of data and an ongoing shift to edge-based locations. The adoption of Industry 4.0 applications in relation to intelligent robotic units is leading to increased factory and warehouse autonomy. As factories integrate these autonomous mobile robots (AMRs), supplanting less intelligent automated guided vehicles (AGVs), there are hundreds of constituent component elements now being involved - and these are producing massive amounts of data on a continuous basis.
 
Since this data must be aggregated and scaled, open platform communications unified architecture (OPC UA) and time sensitive networking (TSN) are becoming more widely adopted. As a machine-to-machine communication protocol, OPC UA provides a secure and platform-independent service-oriented architecture. Furthermore, when combined with the outstanding real-time capabilities of TSN, OPC UA is a key solution to addressing the evolving manufacturing industry - enabling rapid, low-latency responsiveness to be benefited from. Specifically, with OPC UA being overlayed by TSN at the sensor and actuator level, it is possible to achieve continuous insight and data flow without distributing operations or impacting on security. This is something that is essential for delivering scalable enhanced feature sets and system reliability as the digital transformation of production lines continues. As a result, developers need solutions that can implement both OPC UA and TSN frameworks, so as to process data from a variety of different outputs quickly and securely.
 
Increasing data generation is also driving demand for cloud-to-edge networking. Edge processing is the act of extracting and then distilling captured data as close to (or directly on) the device it originated from, thus enabling processing to be conducted in real time. Such rapid processing improves efficiency and reduces the prospect of latency-sensitive bottlenecks occurring, thereby enhancing overall system performance and improving the user experience (UX) derived.
 
Edge computing continues to witness rapid growth, as enterprises seek lower latency plus more secure processing than centralised cloud data centres are able to provide. With billions of edge computing devices being deployed, this rise in edge-based processing requires developers to leverage solutions that are secure, flexible and adaptable, so they can have confidence that data flows will run smoothly. FPGAs are an optimal solution for keeping pace with their evolving needs and ever more exacting demands.
 
How FPGAs can address industrial challenges 
The application of FPGAs can alleviate a myriad of challenges associated with autonomous robotics. The technology can be utilised in numerous ways - ranging from motion control and real-time connectivity all the way to security and machine vision. In addition, there are 3 primary use cases for how FPGAs attend to industrial requirements. These are as follows - high-performance and low-power functionality, parallel processing capabilities, plus deployment flexibility.
 
High-performance and low-power functionality - FPGAs provide elevated levels of performance compared to most other devices, including microcontroller units (MCUs), while consuming only a fraction of the power. They can also handle at least 2x as many motors per device. For developers in the industrial space, this is particularly useful - as it allows them to effectively control multi-axis motors more precisely and efficiently. FPGAs are critical for AMRs that require specific timing, containing hardware-accelerated post-processing that reduces the latency of the overall pipeline and brings benefits to time-sensitive applications. For example, with high-speed automated factory conveyor belts, FPGAs enable machine vision cameras to quickly sense, analyse and act on IIoT data to identify defective products. This empowers developers to optimise supply chain processes and minimise downtime, ultimately leading to higher productivity and profitability.
 
Parallel processing capabilities - The increasing number of sensors and cameras in modern day robots underlines the need for AI inferencing. FPGAs’ parallel processing capabilities aggregate sensor and camera data to quickly analyse high volumes of data/information. While graphics processing units (GPUs) and central processing units (CPUs) can only cope with one batch of data/information at a time, FPGAs are able to deal with multiple functions simultaneously, thereby consolidating functions to achieve more with less power budget. These capabilities also allow FPGAs to monitor multiple communications buses - protecting and detecting, as well as recovering multiple platform firmware elements at the same time. For example, in automotive manufacturing, the bridging functions performed by FPGAs allow multiple high-resolution cameras to have their signals combined or multiplexed together near where the vehicle’s cameras are located, and then have those signals separated again or demultiplexed when they arrive at the system-on-chip (SoC). Equipped with this functionality, developers have scope to work across various connection types, allowing them to bring more data into the pipeline, then quickly and securely transfer it across devices.
 
Deployment flexibility - Manufacturers need flexible solutions that provide deployment flexibility as new AI, ML and IIoT applications are integrated into robotic systems. Especially as factories and warehouses move towards edge topologies, FPGAs are critical for fusing power efficiency with customisation abilities, alongside their reprogrammable nature. FPGAs contain a broad range of highly efficient processing and flexibility attributes at the system level to accommodate the rapidly changing landscape. Unlike custom-built application-specific integrated circuits (ASICs), FPGAs offer the added benefit of being reprogrammable or updatable after the system has been deployed. This provides developers with the unique ability to create devices that can be reconfigured. Additionally, since industrial equipment often has product lifetimes measured in decades (not just years), the prospect of FPGAs updating existing machines in order to support new standards or meet emerging technical demands can prove invaluable.
 
Enabling next generation robots
The rapid integration of Industry 4.0 technologies will result in a reshaping of the manufacturing sector, and autonomous robotics will only continue to become more advanced and interconnected in the years ahead. FPGAs will serve as a pivotal tool for addressing the core challenges involved - bringing innovation, adaptability and efficiency, to keep pace with this accelerated demand for innovation.


Contact Details and Archive...

Print this page | E-mail this page