Cloud-based processing offers extensive resources to support complex artificial intelligence (AI) workloads, like computer vision and voice interaction. But when humans engage directly with machines, high-volume, raw data transfers can increase response latency, thus sub-optimal user experiences.
In contrast, many advanced edge platforms support fast, local AI inference for real-time human-machine interactions. By allowing sensitive data to remain on-device, local processing reduces the security and privacy concerns of transmitting data to the cloud. Moreover, edge processing supports operational continuity in the event of network outages.
Still, by combining edge AI platforms with cloud connectivity, developers can orchestrate data and device management frameworks to improve long-term success. In smart building management, for example, connected, intelligent human-machine interfaces (HMIs) can enhance operational efficiency, safety, and security.
Intelligent HMI use cases for smart building management
In high-footfall settings, building management teams can benefit from AI automation in core safety and security workloads, such as building system control, directing visitors, and crowd monitoring.
Connected HMIs allow personnel to easily visualize smart building metrics like occupancy or localized temperatures, and to remotely control building systems. With edge intelligence, however, easy-access terminals can automate building systems such as local heating and lighting to reduce energy consumption in unoccupied rooms.
Smart lobby screens further increase efficiency by validating self-check-ins using facial recognition and voice interaction AI. These HMIs, connected to centralized management systems, can also provide visitors with personalized information, directions, or access, and notify staff of visitor arrivals, all without sensitive user data exiting the device.
Similarly, low-latency computer vision can recognize unauthorized entry, possible medical emergencies (e.g., a person falling), or other unusual activity within crowded scenes. Upon flagging an incident, connected platforms can deliver visual logs and contextual data, thus providing informed staff assistance.
Such HMI capabilities at the edge require careful consideration in design and infrastructure for long-term deployment success. Public-facing HMIs need robust, ingress-protected touch displays for reliability and ease-of-use. To drive these applications, developers also need sophisticated edge computing platforms, ideally using heterogeneous architectures with powerful CPU and GPU cores alongside dedicated AI acceleration. Industrial connectivity options then enable wider integration across building management systems.
A comprehensive HMI software stack is also required. Developers must select an operating system (OS) allowing easy customization while supporting AI model deployment on edge hardware. Secure update capabilities are a foundation for long-term device health, aided by an overarching cloud layer for data orchestration and fleet lifecycle management.
Creating this combination from scratch places a huge burden on development teams tasked with rapid HMI deployment at-scale. Here, engineers benefit from leveraging ready-to-go platforms that provide a complete, reliable baseline for intelligent HMI development, reducing integration complexity and shortening time-to-market.
A scalable solution for faster HMI development
Already, a smart lobby screen demonstrates this HMI foundation, based on SECO’s production-ready Modular Vision 10.1 MX95 industrial HMI platform running Clea OS. On top of building system control, the HMI provides:
- Custom interactions for approaching visitors or employees, driven by facial recognition and voice interaction AI models, all executed locally.
- Alert management in the event of security, safety, and equipment anomalies, with step-by-step guidance on authorized interactions.
- Crowd monitoring and fall detection to mitigate overcrowding and improve emergency responsiveness through real-time, automated oversight.
Such functionality is made possible by SECO’s highly integrated hardware and software ecosystem. The Modular Vision 10.1 MX95 offers a ready-to-integrate 10.1-inch capacitive touch display with 1280 × 800 resolution and ≥400 nits brightness for clear visibility across indoor environments. This panel-mounted screen is protected by chemically strengthened cover-glass for long-term durability. IP66-rated construction provides a strong resistance against dust and water.
Driving this display is the NXP i.MX 95 multi-core processor, featuring six Arm Cortex-A55 cores for the application domain, two Cortex-M-class cores for real-time and system management tasks, an Arm Mali GPU for 2D and 3D graphics acceleration, and an integrated neural processing unit (NPU) for hardware-accelerated edge AI. Soldered-down LPDDR5-6400 memory supports processing resources, boosting industrial operating reliability.
The rear panel features a range of industrial connectivity options for deployment integration, including serial ports for legacy I/O, USB for peripheral systems, and Ethernet. The Modular Vision 10.1 MX95 accepts a wide 9 to 32 V DC input voltage, further supporting integration with existing industrial infrastructure.
Out of the box, the Modular Vision 10.1 MX95 comes preinstalled with SECO’s Clea OS, a highly modular, customizable Linux image based on Yocto. As well as enabling rapid HMI deployment, the default distribution features built-in security, remote monitoring, and over-the-air (OTA) update mechanisms to support device lifecycle management and cybersecurity commitments. By combining Clea OS with Clea Astarte and Clea Edgehog for data orchestration and fleet management, smart building systems can dramatically increase deployment scalability through centralized device governance.
When integrating real-time edge intelligence with the Modular Vision 10.1 MX95, SECO App Hub eases edge AI development through pre-tested models for workloads like people counting, facial recognition, and more. By presenting engineers with a complete, industry-ready hardware, software, and edge AI development platform for building intelligent HMIs, SECO can be an effective, long-term working partner to developers and system integrators in demanding industrial applications like smart building management.
Conclusion
Real-time local AI presents developers with many opportunities to enhance operational efficiency and data privacy. For smart building applications, it offers staff a force multiplier through intelligent HMIs; automating visitor guidance and providing additional safety and security monitoring through computer vision. SECO’s Modular Vision 10.1 MX95 provides a proven, robust platform for building such HMI deployment at scale; with long-term, evolving edge intelligence supported by Clea OS and the wider Clea software framework.
Visit seco.com to explore how SECO’s Modular Vision platforms can accelerate industrial HMI deployments across a wide range of industries.