Access Official Technical PDF

Share Summary:LinkedInEmail

Industrial automation technologies increasingly depend on connected devices, machine vision and software that can manage data and models throughout their lifecycle. A catalog of Israeli companies presented at CES 2020 offers a snapshot of approaches relevant to automation engineers and technical decision-makers—from securing IoT endpoints to supporting deep-learning development and tracking moving equipment.

The companies described serve a range of markets, not all exclusively industrial. However, several technologies address challenges that also arise in industrial IoT, robotics, vehicle systems and Industry 4.0 environments. Their examples show why evaluating an automation solution involves more than its core algorithm: integration, data handling, communications and operating conditions matter too.

Industrial automation technologies span the full data lifecycle

Deep-learning applications rely on more than a trained model. Teams need to manage data, experiments and processes as they develop and deploy models. Allegro.ai described an enterprise platform for deep-learning lifecycle management that combines experiment and process management with data management and deep-learning operations. The company also stated that data and models remain within the customer’s firewall.

For organizations developing computer-vision applications, this kind of lifecycle approach is relevant to the work surrounding model development and deployment. Allegro.ai identified use cases including autonomous vehicles, drones, security and medical imaging. The catalog also describes support for continuous learning and personalization after deployment, with edge devices able to improve independently over time.

These capabilities do not remove the need for application-specific validation. They do, however, highlight practical questions for engineering teams: how are training data and model versions managed, where does sensitive information reside, and what processes support a deployed model?

Securing connected devices and industrial networks

As connected equipment becomes part of broader IoT and industrial IoT deployments, security at both the device and communications levels becomes important. Firedome described a software agent installed on IoT devices that monitors real-time activity for abnormal behavior. Its architecture also includes a cloud-based AI engine that communicates with the agent, receives data and returns commands. The company presented the agent as lightweight and designed for integration into both new and aftermarket devices.

This device-level approach differs from CipherSiP Systems’ focus on network communication. The company described technology intended to expand bandwidth and add security without changing existing network topology, protocols, messages or standards. Its stated scope includes automotive and industrial networks, with examples such as CAN, CAN-FD/XL, LIN and SCADA. The company also described identifying the sender of messages between electronic control units and controllers.

For integrators, the distinction is useful: endpoint monitoring and communication-layer controls address different parts of a connected system. The catalog does not establish that either approach is suitable for every architecture, so protocol compatibility, deployment constraints and security requirements remain essential evaluation criteria.

Machine vision for variable operating conditions

Vision systems can be limited by poor lighting, weather or visual noise. BrightWay Vision described gated imaging that synchronizes a camera with a pulsed near-infrared illuminator. The camera captures light returning from a selected distance, and repeated measurements form an image. According to the company profile, this approach is intended to reduce backscatter from nearby particles and produce an image of a specific range.

TriEye presented a different approach: a shortwave infrared camera based on a CMOS sensor design using nanophotonics. The company positioned SWIR sensing for low-visibility and nighttime conditions and described potential use in advanced driver-assistance systems and autonomous vehicles. Its profile also noted remote material sensing for detecting hazards such as black ice.

These examples concern perception in demanding conditions rather than factory inspection specifically. Still, they underscore considerations relevant to automation projects that use cameras: the sensing method, environmental conditions and the type of information required all affect system design. A camera that improves visibility or distinguishes material properties may support a different task from a conventional image sensor.

Tracking and communication support automation infrastructure

Sixdof Space described an optical tracking system that uses existing room lights as location beacons, without modifying them. The company said its technology reports position and orientation—six degrees of freedom—to a host system, and identified potential applications including robots and warehouse vehicles. For automation teams, accurate spatial tracking can be a building block in systems that coordinate mobile equipment or other tracked assets, subject to the requirements of the specific application.

Across these profiles, a recurring theme is integration into existing environments. CipherSiP emphasized maintaining network protocols and topology; Sixdof Space described using existing lighting; and Firedome presented software intended for aftermarket IoT devices. Such approaches may be relevant where changing installed infrastructure is difficult, but the catalog does not provide a basis for comparing implementation effort or performance across vendors.

What automation teams can take away

The CES 2020 company profiles illustrate several areas to assess when planning connected automation systems:

  • Model operations: Consider how data, experiments and deployed models will be managed, and where those assets are stored.
  • Security layers: Evaluate protections for both individual devices and communications between controllers and networked equipment.
  • Perception conditions: Match sensing technologies to the lighting, weather and detection challenges in the intended environment.
  • Integration needs: Check how a solution interacts with existing protocols, infrastructure and installed devices.
  • Application fit: Treat vendor-described use cases as starting points for technical assessment, not as substitutes for validation in the target system.

For industrial automation leaders, the practical lesson is that innovation is distributed across the system: sensing, model management, device security, communications and integration all influence what a deployment can achieve. These CES 2020 examples provide a useful historical snapshot of technologies addressing those challenges. Any adoption decision should be grounded in the plant’s architecture, operating conditions and security requirements.

Access Official Technical PDF

About The Author