

Artificial Intelligence Solutions for the Network Edge
While developers may choose to perform AI inferencing in the cloud, where hardware resources are hardly constrained, there are often good reasons for embedded systems to perform inferencing locally at the network edge, including:
- Lower latency
- Simpler hardware with less maintenance
- Immunity to poor network connections
- No costs associated with network data transfer
- Security against malware and other network-borne attacks
Local inferencing may be performed on a range of devices which embedded developers are already familiar, including microcontrollers, microprocessors/SOMs/SBCs and FPGAs. Through analysis of time-series sensor data and utilization of neural network models, they enable a broad set of applications such as the detection or identification of objects, gestures, keyword phrases, and industrial system anomalies.
Product Families
AI edge solutions are available from multiple manufacturers of IC products, modules/boards and development tools. Demonstrations are available today for the following manufacturers and product families:
Description | Manufacturers | |
Processor solution with built-in peripherals for event-based applications | Maxim Integrated, NXP, Renesas, STMicroelectronics | |
Advanced processor solution with more/faster cores and interfaces, larger memory, virtual machine support and ability to run Linux/Windows/Android | NXP, Renesas, STMicroelectronics | |
Parallel processing solution with large number of high-speed/customizable interfaces and optional soft or hard processing cores | Lattice Semiconductor, Microchip (Microsemi), QuickLogic | |
SOMs/SBCs | Complete system on module or single board computer solution with microprocessor/FPGA, memory and power management for faster time to market | AAEON, Advantech, Compulab, iWave Systems, Ka-Ro, Technexion |
Development Tools | Integrated HW/SW solution for SoC platform or Cortex-M class microcontrollers to streamline collection and labeling of training data, optimize algorithms and validate code without data science expertise | QuickLogic |
Solutions by Application
A list of applications and products that have been demonstrated to address them are included in the table below, where you can follow links to view demos and check stock. Choice of product will depend on specific application requirements including speed and accuracy.
Microcontrollers & Tools | Microprocessors | FPGAs/SoCs | |
Predictive Maintenance / Condition Monitoring / Anomaly Detection |
| ||
Activity Recognition |
| ||
Gesture Detection |
|
| |
Sound Analysis |
| ||
Key Phrase Detection / Voice Control | |||
Face Expression Recognition |
| ||
Face Detection |
| ||
Face or Object Tracking | |||
Traffic Monitoring | |||
Human or Object Detection | |||
Human or Object Counting | |||
Pose Detection |
| ||
Image or Object Classification / Recognition | |||
Facial Recognition / Identification |
Future Electronics Development Kits
In addition to manufacturer board platforms available for the individual devices listed above, two additional kits designed by Future Electronics System Design Centers (SDC) may be used in the development of your AI application.
Avalanche Development Kit![]() | Platform with Microchip MPF300TS 300kLE PolarFire FPGA and VSC8531 Gigabit Ethernet PHY, Panasonic PAN9420 WiFi module, and Arduino/mikroBUS/PMod connectors, ideal for object detection and monitoring in smart embedded vision applications. |
Compagno Development Board![]() | STWIN development kit which provides training and valuable information about Machine Learning (ML). Designers can perform cloud air quality demos with Scriptr using this board. |
Need Technical Assistance?
Future Electronics has a global team of specialist engineers with decades of experience in microprocessor and FPGA embedded processing applications. They are available to assist you in selecting the best solution for your AI application. Tell us about your project using the form below, and a member of our technical team will contact you.