A New Direction for Industrial Automation
Seeing Machines has secured an advanced development contract with a global industrial technology company specialising in factory automation and human-robot interaction. The initial programme will focus on developing a robotics proof-of-concept for a factory environment, with the potential to support broader collaboration in industrial automation and robotics.
From an industrial automation perspective, the importance of this agreement extends beyond a single proof-of-concept. It represents an interesting shift from conventional machine vision toward human-aware automation, where machines are expected not only to detect physical objects but also to understand human behaviour and operational context.
From Driver Monitoring to Factory Perception
Seeing Machines has built its reputation primarily through vision-based monitoring technologies used in automotive, commercial fleets, off-road applications and aviation. Its systems use AI, embedded processing and optical technologies to analyse human attention, gaze direction and behavioural indicators associated with operational risk.
The company's Perception Map technology provides a foundation for extending these capabilities into robotics. Instead of asking only where an object is located, an industrial perception system can potentially determine where a worker is looking, how the worker is positioned and whether their behaviour indicates an intention to interact with a machine.
This is a significant conceptual step for factory automation. Traditional machine vision is generally centred on objects, positions and processes, while human-centred perception introduces another variable: the human operator's state and intent.
Human-Robot Interaction Becomes a Core Automation Challenge
As collaborative robots, autonomous mobile robots and semi-autonomous production equipment become increasingly common, people and machines are sharing the same operational space more frequently.
In these environments, conventional safety functions such as fixed barriers, safety scanners, light curtains and emergency stops remain essential. However, they do not necessarily provide the contextual understanding required for more sophisticated human-machine interaction.
A perception system capable of identifying human position, attention and behavioural cues could complement traditional safety architecture. For example, a robot could potentially distinguish between a worker who is simply passing through an area and one who is actively preparing to interact with the robot.
In my view, this is where the industrial value of technologies such as Seeing Machines' platform could become particularly interesting. The next generation of factory automation will not be defined solely by how accurately machines perceive objects, but by how intelligently they interpret people.
Proof-of-Concept Is the Critical First Step
The current contract is an initial development programme rather than a confirmed large-scale production deployment. Seeing Machines has not disclosed the identity of the industrial technology company or the financial terms of the agreement.
Technically, this makes the proof-of-concept phase especially important. The real challenge will be demonstrating that technology developed for transport environments can operate reliably under factory conditions.
Factories introduce different variables, including variable lighting, occlusion, multiple workers, protective equipment, rapidly changing movements and complex interactions between operators and machinery. Industrial systems must also meet demanding requirements for availability, deterministic behaviour and integration with existing control and safety architectures.
Therefore, successful commercialisation will depend not simply on recognition accuracy, but on how effectively the perception technology can be integrated into a complete automation system.
Beyond Machine Vision: Understanding Human Intent
The most interesting aspect of this development is the potential transition from human detection to human understanding.
Conventional industrial vision systems can already identify workers, components and movement with impressive accuracy. The next challenge is determining what those observations mean in the context of a manufacturing process.
For example, identifying that a worker has entered a robot cell is useful. Understanding that the worker is approaching a specific machine, looking toward a particular operating area and preparing to perform a manual intervention could provide much richer information for an intelligent control system.
This could eventually support more adaptive automation strategies, such as dynamically adjusting robot behaviour, modifying operating speed or providing contextual assistance to operators.
However, these capabilities should complement rather than replace certified industrial safety functions. AI-based perception should be treated as an additional intelligence layer, while deterministic safety systems remain responsible for safety-related protective actions where required.
Potential Impact on Smart Factory Architecture
The agreement also highlights a broader evolution in industrial automation architecture. Historically, automation systems have been built around PLCs, motion controllers, robots, sensors and industrial networks. Increasingly, AI-based perception is becoming another layer within this architecture.
A future smart factory may therefore combine:
-
Traditional PLC and DCS control systems
-
Safety PLCs and certified protective devices
-
Industrial robots and collaborative robots
-
2D and 3D machine vision
-
AI-based human perception
-
Edge computing and embedded AI
-
Industrial Ethernet and real-time communication
-
Higher-level manufacturing and analytics platforms
The key engineering challenge will be integrating these technologies without compromising determinism, cybersecurity, functional safety or system maintainability.
A Strategic Opportunity for Seeing Machines
For Seeing Machines, the factory robotics programme offers a potential route into a market that is structurally different from its established transportation business.
The company's more than 25 years of experience in human factors, attention monitoring and risk assessment provides a technical foundation for this expansion. Nevertheless, industrial automation has its own certification, integration and lifecycle requirements, so success in transportation does not automatically guarantee success in factory environments.
The initial proof-of-concept should therefore be viewed as a technology and market validation exercise. If it develops into repeat engineering programmes, production deployments or integration into commercial robotics platforms, the opportunity could become considerably more significant.
My View: Human-Centred Automation Is the Next Logical Step
I believe the most important message behind this contract is not the robotics application itself, but the direction of industrial automation.
Factories are becoming increasingly autonomous, yet humans remain central to maintenance, quality control, material handling, troubleshooting and production decision-making. Completely separating people from automated equipment is not always practical or economically attractive.
The more realistic path is to make machines better at understanding the people working around them.
That requires a combination of machine vision, AI, real-time processing, industrial controls and safety engineering. Seeing Machines' move into factory robotics is therefore worth watching—not because a single proof-of-concept immediately changes the automation market, but because it illustrates how human perception could become an increasingly important component of future industrial control systems.
Conclusion
Seeing Machines' new robotics development contract represents a meaningful expansion of its Human-Centred AI strategy into industrial automation. The initial project remains at the proof-of-concept stage, and its commercial outcome is not yet known.
Nevertheless, the technology direction is compelling. As factories become more collaborative and autonomous, the ability of machines to perceive, interpret and respond appropriately to human behaviour could become an important differentiator in next-generation automation systems.
