CAD | USD

Robotics Is Moving Off the Maker’s Bench and into the Mainstream

By European Editors

Contributed By Digi-Key's European Editors

Getting started in robotics has never been easier or more rewarding. Research carried out today could lead to commercial success tomorrow. With a wide range of open source projects now available, engineers have the opportunity to focus on control systems that offer a better ‘plug and play’ solution.

Not so long ago, robotics was the domain of either enthusiastic hobbyists or large manufacturers in the industrial automation industry, with a large void between them. Today, that void is closing and as these two extremes collide, the definition of what a robot is and isn’t good for is changing.

Robotics is a rapidly expanding area of research, but more than that it now represents a large opportunity for enthusiasts, industrialists and everyone in between. Applications — not limitations — will shape the robotics industry and, in truth, the potential applications for robotics isn’t limited to monotonous and repetitive tasks such as welding a car frame, or a highly specialized but restricted process like suturing a patient. In reality, limiting robotics to only those applications so far imagined underplays its potential and, as some may believe, its necessity in tomorrow’s world.

Robots are evolving; they may never be self-aware but the same can be said of most of Earth’s species. Their evolution may always be limited by man’s ability to imagine and develop new technologies, but man’s imagination is a reliable source of constant surprise.

Enabling technologies

The tendency to emulate the human form when imagining robots isn’t merely an affectation, it is logical to assume that robots will need to function in the world as it exists, not as an engineer might design it. This has led to the development of advanced and (some may argue) intelligent vision systems using stereoscopic cameras to judge depth, size, distance and speed.

Sensors in general (not just for vision) are a crucial area of research and development for the robotics industry. Textiles that react like skin (for detecting pressure, heat and proximity) may one day coat the limbs of a robotic form. Imagine how much simpler it would be for a robot to walk if it could detect and comprehend the floor beneath it in the same way ambulatory animals do. Extending this sensitivity through a ‘central nervous system’ would require further sensors at every joint and muscle; sensor technologies that are undoubtedly being researched.

Processing this huge amount of data would, of course, require a large and powerful processing engine. Here, the future is perhaps easier to imagine, as the semiconductor industry is already producing integrated circuits with features in the very deep sub-micron range, along with innovations such as FinFETs and graphene. Transistor density, as dictated by Moore’s Law, is expected to continue to increase for many years to come, allowing ever-greater processing power in ever-smaller physical dimensions.

All of this processing power will be put to good use running software that delivers artificial intelligence. There has been a corresponding resurgence in interest in neural networks, thanks in large part to more powerful multicore processors that are able to handle their complexity. This will see deep learning algorithms being used extensively in robotics.

The ability to design and manufacture complex and intricate parts using 3D printing techniques is also fuelling the robotic revolution; an example is the Open Robot Hardware initiative, which hosts a number of projects featuring open source mechanical and electrical components. It is now possible to download the design files for parts of a robot, print them at home and construct a robot using open source hardware and software for control purposes.

More than maker

There is a significant amount of development taking place in the ‘maker’ sphere, where open source projects gain much of their traction. The number and diversity of open source robotic projects now available extends beyond the maker community and is largely driven by colleges and universities, with projects that allow the study of robotics using low cost and often 3D-printed kits.

While open source hardware is becoming more widespread, open source software has a much longer history and higher market penetration, particularly within industrial and consumer applications (Linux and Android being the two most prominent). This now also includes the Robot Operating System (ROS) project, which, rather than being an actual ‘operating system’ as most people might think of one, is a modular framework comprising middleware libraries that engineers can use to create robot software. It runs on the Linux derivatives Ubuntu and Debian, so requires a host processor capable of running Linux. It can also be installed directly to a number of preconfigured robots (visit ROS Kinetic installation instructions for more details).

There are a number of sub-projects within ROS, including ROS-Industrial, which extends ROS to the manufacturing domain. ROS is also working on a real-time implementation of the framework, called ROS 2. This would include elements that can be ported to microcontrollers which might be used to control the actions of parts of a robot, for example, allowing a more distributed approach to design.

The fusion between microcontrollers and microprocessors means that some higher-end devices targeting microcontroller-like applications are actually microprocessors able to run Linux-based operating systems. The main enabler in this area has been ARM’s family of Cortex cores, ranging from the Cortex-A (application), to the Cortex-R (real-time), and of course the Cortex-M (microcontroller). The most prominent of these cores in the embedded space is still the -M family, but devices using the -A family are now present in devices aimed at high-end industrial and automotive applications, as well as wired/wireless communications. A good example of this is the Zynq-7000 SoC family, which integrates a dual-core ARM® Cortex®-A9 MPCore with up to 444k logic cells. The closely integrated nature of the FPGA fabric and dual-core processor makes the Zynq-7000 SoC suitable for applications that require high levels of parallel processing, such as advanced imaging in machine vision systems. The integration of the ARM Cortex-A9 cores also means it is capable of running the ROS framework in a commercial/industrial robot.

The architecture of the Zynq-7000 SoC made it the ideal choice for the vision system of a swarm of robots developed by Three Byte Intermedia for the Museum of Mathematics (MoMATH), demonstrated at Embedded World 2015.

Diagram of Xilinx Zynq-7000 architecture

Figure 1: Xilinx Zynq-7000 architecture.

Worlds in motion

Motion lies at the heart of robotics, whether that is full ambulatory action or a device with several degrees of freedom for moving ‘things’ around in a defined area. Motor control is fundamental to motion in robotics, which encompasses both Brushless DC motors (BLDC) for controlled angular motion and AC motors for automation and transportation. While BLDCs are credited as being crucial in the robotics industry, motor control in general will remain key. Modules that incorporate DC motors with reducing gears and controllers are now widely available and used within robotics. These modules typically use a relatively simple serial protocol for bidirectional communications, allowing many modules to be controlled by a single MCU. Microcontrollers that also incorporate AC motor control and the ability to interface to multiple sensors could provide a compelling solution to robotic control systems.

Industrial microcontrollers, like the XMC4500 series from Infineon have been optimized to offer a combination of motor control, sensor interfacing and connectivity. The XMC4000 family is based on the ARM Cortex-M4 with floating point and DSP instructions. This gives it the processing power needed to execute the complex algorithms used in robotic applications, with support for the protocols used in industrial environments such as CAN, as well as Ethernet, USB and SPI. It also has an SD and Multi-Media SD card interface and an interface dedicated to communicating with external memories and off-chip peripherals. Furthermore, it features an LED and Touch Sense module, which is used to drive LEDs and control a touch pad when a Human-Machine Interface is needed.

Block diagram of the Infineon XMC4500

Figure 2: Block diagram of the Infineon XMC4500.

The ability to process MAC (multiply and accumulate) instructions efficiently is crucial in real-time control systems like robotics, and a feature of the Cortex-M4’s DSP instructions, but it isn’t the only choice in microcontroller architectures in this field. The 32-bit RX core from Renesas, as used in its RX200 series, uses a variable length CISC Harvard architecture with a 5-stage pipeline, whose accumulator can handle 64-bit results in a single instruction, as well as single-cycle multiplication instructions. The RX210 group of MCUs delivers 78 DMIPS at 50 MHz and offers a wide range of peripherals and up to twenty extended-function timers. The RX210 also integrates hardware features intended to make compliance with IEC 60730 (Safety Standard for Household Appliances) easier to achieve; as robots will likely be used in a domestic environment, this could offer a major benefit.

Block diagram of the Renesas RX210

Figure 3: Block diagram of the Renesas RX210.

Conclusion

Robotics is an emerging and rapidly evolving area of engineering, which has only just begun to explore its potential. Fears of creating sentient beings bent on annihilating mankind, as often depicted in science fiction, are unlikely to hold back research. If you would like to start exploring the world of robotics, why not take a look at the kits available and be part of the future, today!

Disclaimer: The opinions, beliefs, and viewpoints expressed by the various authors and/or forum participants on this website do not necessarily reflect the opinions, beliefs, and viewpoints of Digi-Key Electronics or official policies of Digi-Key Electronics.

About this author

European Editors

About this publisher

Digi-Key's European Editors