Use an SBC to Quickly Implement Edge AI in New or Retrofit Applications

Developers of Internet of Things (IoT), robotics, computer vision, and industrial applications face growing pressure to embed intelligence into their highly connected edge designs. For teams working under tight deadlines, this pressure extends beyond application software development. Selecting hardware capable of running high-level operating systems such as Linux alongside deterministic real-time functions is challenging enough, but when intelligence is retrofitted into existing infrastructure, such as in industrial automation and smart building applications, additional platform suitability requirements arise.

What developers need is a familiar, proven, flexible, and capable platform to quickly prototype and develop production-ready designs.

This article discusses the challenges developers face with processing and retrofit projects at the network edge. It then shows how an Arduino single-board computer (SBC) can be used to address these challenges.

Building edge intelligence under strict resource constraints

Edge intelligence encompasses artificial intelligence (AI) inference and decision-making, running on a local platform. Key advantages of edge-based intelligence include reduced reliance on always-on connectivity, improved privacy and security, and ultra-low latency, all of which benefit designers of robotic and industrial safety systems.

For robotic devices, edge intelligence enables real-time motion control, obstacle avoidance, and adaptive behavior, delivering the deterministic response times critical to autonomous operation. For industrial safety systems, edge intelligence enables immediate hazard detection, predictive maintenance, and rapid shutdowns, minimizing equipment damage and worker risk. Overall, edge intelligence provides the responsiveness, resilience, and reliability required for real-time AI applications.

But limited hardware resources impose significant constraints. Cloud-based systems can scale as needed, whereas edge-based intelligence must balance onboard processing against power envelopes and thermal constraints. Real-time AI workloads such as computer vision, sensor fusion, and robotic control can saturate processing resources, increasing power consumption and heat generation. Excessive thermal load on a processor can lead to reduced inference performance, system instability, or thermal throttling, in which the processor automatically slows down to cool off when it gets too hot.

Power envelope limitations are equally critical when edge systems operate on batteries, mobile power systems, or otherwise restricted power supplies, where energy efficiency directly affects runtime and reliability. Retrofitting often introduces challenges. Existing platforms typically have limited space, making it difficult to add AI accelerators, cooling systems, or additional memory. Legacy systems might have outdated or proprietary interfaces that require adapters or custom integration to connect modern hardware to existing technology.

The advantage of SBCs for rapid development and retrofit success

Edge intelligence can be implemented using an SBC. These compact, embedded computing platforms integrate a processor, memory, storage interfaces, networking, and peripheral connectivity on a single printed circuit board (pc board), making them ideal for edge applications.

SBC platforms can provide out-of-the-box functionality, including an operating system (OS), networking stacks, camera interfaces, storage, and hardware acceleration, without requiring a custom board design. SBCs typically include verified I/O and pre-validated interfaces that are supported by existing software ecosystems. In general, they offer broader capabilities than standalone embedded controllers and often run open-source Linux.

For retrofits, an SBC paired with real-time processing resources can reliably interact with legacy sensors, actuators, and interfaces while supporting predictable timing for scheduling, processing, or task response. Retrofits can also benefit tremendously from intelligent platforms, though they impose demanding requirements for intelligent edge applications, including compact physical dimensions, high performance per watt (PPW) in heterogeneous computing architectures, and low thermal design power (TDP).

Heterogeneous computing pairs specialized processors with the tasks they handle best. This relieves the bottlenecks in traditional architectures and enables real-time responsiveness, power conservation, and high performance in constrained environments. High PPW supports a constrained power budget, as AI inference may run in parallel with sensor fusion in real time. Low TDP matters because excessive heat reduces reliability, increases the risk of hardware damage, or can trigger thermal throttling.

SBCs can make it easier to accommodate all these retrofit requirements. For IoT and industrial applications, comprehensive coverage of wired and wireless connectivity options can reduce system expansion requirements and simplify integration. For robotics and computer vision systems, high-speed image signal processing and AI inference can enable real-time navigation, object detection, and autonomous decision-making with deterministic response characteristics.

A dual-architecture SBC with a familiar form factor

Microcontroller units (MCUs) provide excellent real-time hardware control, so many SBC designers pair them with powerful CPUs in a dual architecture. For example, Arduino’s UNO Q (Figure 1) is an SBC that combines Qualcomm's Dragonwing QRB2210 processor (running a full Debian Linux OS with upstream support) with the real-time responsiveness of a dedicated STMicroelectronics STM32U585 MCU (running Arduino sketches on Zephyr OS). The power-efficient Dragonwing QRB2210 features a quad-core Arm Cortex-A53 CPU (up to 2.0 gigahertz (GHz)) with an Adreno graphics processing unit (GPU). AI inference can run on the GPU and CPU.

Figure 1: The UNO Q board integrates a CPU, GPU, and real-time MCU within a single development environment. (Image source: Arduino)

Dual, 13 megapixel (MP), 30 frames per second (fps) image signal processors (ISPs) support image pipeline processing and are suitable for advanced features such as machine vision. An integrated digital signal processor (DSP) subsystem further supports lightweight AI inference and multimedia processing. Fanless operation simplifies passive thermal management.

Two versions of the Arduino UNO Q are compatible with traditional Arduino shields and accessories. The ABX00162, with 2 gigabytes (Gbytes) of RAM and a 16 Gbyte embedded MultiMediaCard (eMMC), is best suited for cost-efficient, dedicated, lightweight applications. The ABX00173, with 4 Gbytes of RAM and a 32 Gbyte eMMC, supports computer vision, graphics, and edge processing applications. The latter can provide a responsive standalone desktop experience, multitasking with high-level processes, complex AI model support, or expanded local storage.

In a typical Arduino design, the boards include built-in user-controllable RGB LEDs and an 8 × 13 blue LED matrix for rapid status indication. Wi-Fi 5 and Bluetooth 5.1 provide connectivity for remote control or data monitoring. A 7 V_DC to 24 V_DC power input (via a dedicated pin, rather than the USB-C port) enables direct interfacing with legacy power rails without DC-DC step-down converters.

The Arduino App Lab unifies the “dual-brain architecture” of the SBC and MCU in a single, specialized integrated development environment (IDE). This streamlined approach allows engineers to develop, manage, and deploy both SBC and MCU applications in a consistent workflow, reducing complexity for users.

Accelerating UNO Q development through solderless modular expansion

Arduino UNO Q retains the familiar Arduino UNO form factor and header arrangement while extending the platform for intelligent edge and industrial applications (Figure 2). Traditional UNO-compatible pin headers expose GPIO, power rails, SPI, I²C, and UART connectivity, preserving compatibility with existing Arduino shields and embedded workflows.

Figure 2: The Arduino UNO Q SBC retains the standard Arduino UNO form factor and classic headers while adding modern, high-speed connectors. (Image source: Arduino)

High-speed bottom connectors extend capabilities beyond legacy MCU interfaces by enabling faster peripheral communication and higher-bandwidth system integration in compact configurations. The ASX00073 Ethernet PHY shield (Figure 3) demonstrates straightforward expansion into brownfield infrastructure using existing wiring. Similar shield options support motor control and fieldbus communication (such as CAN bus or RS-485) and enable additional capabilities through rapid hardware expansion.

Figure 3: The ASX00073 Ethernet PHY shield simplifies expansion into brownfield infrastructure using existing wiring. (Image source: Arduino)

The Arduino UNO Q also incorporates an onboard Qwiic connector that allows the connection of sensors, actuators, and Arduino Modulino modules via I²C for rapid, solderless expansion, significantly accelerating prototyping and human-machine interface development. For example, the ABX00110 Modulino button board (Figure 4, left) supports operator input and event triggering for industrial control panels and IoT devices; the ABX00107 rotary encoder board (Figure 4, center) enables intuitive menu navigation, parameter tuning, and robotics control interfaces; and the ABX00102 time-of-flight (ToF) board (Figure 4, right) provides proximity sensing useful for robotics, occupancy detection, and safety systems in automation.

Figure 4: Shown are the ABX00110 button board (left), ABX00107 rotary encoder board (center), and ABX00102 ToF board (right). (Image source: Arduino)

Conclusion

As edge AI applications continue to expand across industrial automation, robotics, and the IoT, flexible SBC platforms offer a fast, effective path to reliable, responsive, and intelligent systems. By integrating high-level processing, real-time control, flexible connectivity, and modular expansion into compact, power-efficient hardware, the Arduino UNO Q exemplifies how an SBC can simplify prototyping and modernize existing infrastructure without extensive redesign.