Evaluating the Dragonwing IQ-9075 EVK: Industrial-grade Solutions for IoT, Edge AI and Robotics.

Fabian Muñoz
6 days ago
4 min read

As embedded systems and IoT solutions continue to adopt AI-driven capabilities, the demand for powerful and energy-efficient platforms continues to grow.. The Dragonwing IQ-9075 EVK board maximizes the features of the Dragonwing IQ-9075 processor, optimized for industrial IoT, Edge AI, and robotics applications, being able to sustain through demanding workloads and offering a high-performance power efficient computing power, on-device AI with a dedicated NPU, and a camera and computing processing GPU.

This blog will aim to showcase RidgeRun's platform exploration to give an insight on the features available in hopes of giving a clearer picture to evaluate if the Dragonwing IQ-9075 adjusts to your needs.

Descriptive image of the available features for the Dragonwing IQ-9075

Industry-leading performance, built-in security features and design flexibility in a single board.

The Dragonwing IQ-9075 processor belongs to the IQ9 Series, Qualcomm's family of processors designed to deliver industry-leading performance for AI industrial applications with on-board AI, industrial-grade safety features as well as error correction code (ECC) memory, and developer-friendly design with multi-OS support for Qualcomm Linux and Ubuntu images. The Dragonwing IQ-9075 EVK provides a wide range of AI and embedded use cases, including:

AI Model development and Prototyping: Create, test, and optimize AI models for computer smart vision, language processing, and machine learning applications.
Robotics and Industrial Automation: Accelerate the development of autonomous and intelligent robotic solutions, enabling real-time AI inference and autonomous decision-making at the edge.
Multi-sensor and Camera Fusion: Create intelligent systems that combine data from multiple cameras and sensors to enable capabilities such as object detection, navigation, surveillance, predictive maintenance, quality inspection, and industrial process monitoring.
High Performance Edge Computing: Enable the development of edge AI solutions for use cases in Industrial IoT, smart cities, retail, workplace safety, intelligent buildings, and healthcare, delivering low-latency processing.

Key Hardware Designs:

On-Device Dedicated Neural Processor Unit (NPU)

One of the key strengths of the Dragonwing IQ-9075 is its dedicated Neural Processor Unit, Qualcomm Hexagon DSP with its Dual Hexagon Tensor processors, capable of delivering up to 100 Dense TOPS of performances that runs 13-billion parameter models and generates 12 tokens per second. The RidgeRun team has an in-depth guide on how to enable the unit, run AI sample applications and benchmark efficiency. Customization for its Qualcomm AI Runtime SDK components is also offered through a Yocto-based development. The main benefits of having a dedicated unit for neural processing are:

Offload AI inference from the CPU/GPU resources.
Enhance performance while maintaining power efficiency.
Enables real-time, low-latency AI processing.
Specialized hardware accelerates neural operations.

Industrial-grade security features

The Dragonwing IQ-9075 offers hardware and software solutions to provide protection to user data. It includes hardware-accelerated cryptographic services, a filesystem-level encryption using hardware-wrapped keys, and its own dedicated secure storage technology. It features hardware-assisted virtualization to support multiple isolated operating system environments and secure workload separation. This provides a secure environment for development and ensures a safe environment for user data through secure services, trusted environments and cryptographic solutions. The Dragonwing IQ-9075 Technical Guide, developed by the RidgeRun team, offers an extensive step-by-step on how to enable, customize, and implement security features such as:

UEFI Secure Boot
Trusted Execution Environments (TEE)
Secure File System
Storage Encryption

Hardware and Software Accelerated Multimedia Support

The Qualcomm Dragonwing IQ-9075 platform provides a multimedia infrastructure for developing high-performance embedded and edge applications. Through a combination of hardware-accelerated processing engines, multimedia frameworks, and standard streaming protocols, developers can build efficient video pipelines that capture, process, encode, decode, transform, and transmit multimedia data with minimal CPU overhead. It concurrently process 4K video streams in applications such as computer vision, with 4 MIPI CSI ports supported by the Qualcomm Dragonwing IQ-9075 EVK. The Qualcomm Dragonwing IQ-9075 EVK supports a range of peripherals through its low and high speed connectors, standard interfaces, and communication protocol support. It supports a wide range of multimedia frameworks, such as GStreamer, facilitating development for applications heavily focused on video or audio data processing.

Dev-friendly environment and Software Flexibility

The RidgeRun guide provides comprehensive documentation for the Qualcomm Dragonwing platform, enabling developers to quickly set up and customize their development environment through:

Step-by-step Setup: Extensive instructions for downloading, building, flashing, and configuring Ubuntu images and Qualcomm Linux (Yocto) builds for agile deployment, with detailed customization options and direct control over system components
Multi-OS Support: Flexibility to develop, test, and deploy applications across multiple OS images within their own secure environment.
Long Term Support: The Dragonwing IQ-9075 processor is part of Qualcomm's Product Longevity Program, providing long-term availability and support through 2038 for products requiring extended lifecycle maintenance.
Framework Integration: Seamless integration of multimedia and machine learning frameworks, leveraging dedicated hardware accelerators such as the NPU, GPU, and DSP to efficiently handle AI inference, multimedia processing, and other compute-intensive workloads.

Why the Dragonwing IQ-9075?

The board's hardware and software design is aimed to bring solutions to the following applications:

Edge AI Boxes
Industrial Robots
Machine Vision
Industrial Personal Computers
Autonomous Mobile Robots
Multimedia Processing and Data Handling
AI-Enabled IoT Solutions
Intelligent Surveillance

The platform stands out for its dedicated hardware accelerators, delivering high performance, power efficiency, and optimized execution for AI, multimedia, and industrial applications while providing a developer-friendly environment.