FAST-TRACKING AGENTIC AI: Achieving more effective AI training with synthetic data. p. 12
WHAT IS CANADA’S LEVEL OF ROBOTICS READINESS?
The need for a national robotics strategy from the federal government. p. 16
CLICK PLUS CPUs offer the utmost versatility with up to two option slots for custom control configurations.
With the new Node-RED® option slot module, you can combine the most affordable and user-friendly PLC in the industry with one of the most popular open-source software tools in the industrial sector for developing IIoT applications.
• Create interactive and customer-specific HMI dashboards that can be accessed with any browser
• Easily integrate plant-floor data into existing business systems/databases, e.g., MES or ERP systems
• Access cloud-based services or website APIs for a wealth of information (weather reports, energy/raw material costs, etc.) that can be used to optimize your processes What can it do?
The C2-OPCUA module, when added to a CLICK PLUS CPU, allows the simplicity and reliability of CLICK PLCs to be deployed in many more applications, providing one of the lowest cost OPC UA systems available.
What does it offer?
• Scalability - capable of large amounts of data transfer no matter the application
• Versatility - with OPC UA support, the CLICK PLUS can work seamlessly with numerous 3rd party control platforms, business/IT systems, cloud networks, SCADA systems, etc.
• Enhanced security - featuring authentication, authorization, and encryption methods
• Better efficiency - real-time data exchange for process optimization
shipping does not apply to items requiring LTL transport, but those shipments can take advantage of our negotiated super-low at rates (based on weight) that include brokerage fees. See Web site for details and restrictions at: www.automationdirect.com/canada
Tariff
Steve Wardell, MA’s Lifetime Achievement Award winner, on his 36-year career in the industry and the evolutions he’s seen in machine vision.
By Jared Dodds
How companies are achieving faster and more effective AI training through the use of synthetic data.
By Treena Hein
Outlining the need for a national robotics strategy from the federal government.
By Jared Dodds
Highlighting
adaptive control systems
learn from their environment.
By Jacob Stoller
FROM THE
BY JARED DODDS
From reaction to action
As summer comes to an end, days are becoming shorter and blissfully cooler, vacations are wrapping up and students and teachers are heading back to school.
I find it helpful at the end of each quarter to look back and analyze my writing, particularly the day-to-day news, and see what the trend of the season was. From the beginning of June to the end of August I wrote more than 80 news stories for Manufacturing AUTOMATION, and the topic that stood out to me was clear: the federal and provincial governments.
Whether it was the ongoing trade negotiations at the federal level, including the recent removal of retaliatory tariffs on U.S. goods covered by the Canada-United States-Mexico agreement, the government of Ontario providing funding to multiple communities through the Skills Development Fund Training Stream and investing in the new Protect Ontario Workers Employment Response Centres or premiers from across the country signing agreements to unlock free trade between the provinces, there has been a steady stream of news stemming from both branches of government.
even saw Premier Ford threaten to cut off electricity from the United States. We were constantly responding. That trend has continued into the summer with the previously highlighted funding opportunities and new agreements. These were and remain necessary evils as the manufacturing industry and the country as a whole continues to suffer the effects of tariffs.
But the trade war rages on, and my hope for the first budget tabled by the federal government is a shift in approach — a change from reaction to action.
My hope for the first budget tabled by the federal government is a shift in approach — a change from reaction to action
I took some time to read each story, not only to assess my coverage, but to come to a determination of what each piece of news means when viewed as a collective rather than a one-off. I landed on this realization: the first three quarters of 2025 have been characterized by reactive decision making.
When U.S. President Donald Trump began the trade war in February 2025 by signing the initial tariff orders, we as a country had to scramble. We imposed retaliatory tariffs, removed U.S. alcohol from our shelves and
The signs are there. Canada is building deeper ties with Germany and the EU at large, we have resumed free trade agreement negotiations with a number of South American countries and with multiple Indo-Pacific partners. But the focus on the manufacturing industry, which accounts for approximately 28 per cent of our GDP, needs to be more concrete.
On page 16, I wrote about one area I would like to see change, outlining the importance of a dedicated federal robotics strategy to the manufacturing sector. This style of top-down leadership will build trust with business owners and give them the support they need to integrate new technology.
Being proactive and leading from the front is essential to increase the robustness of the manufacturing industry. Between trade wars, labour shortages, the productivity gap and unseen disruptions, like a global pandemic, Canada and the manufacturing industry have been forced to play catch up for too long. Change is required. Let’s see if the government is ready to not just catch the ball, but run with it. | MA
JIM BERETTA, President, Customer Attraction and host of The Robot Industry Podcast
JONATHAN GROSS, Managaing Director, Pemeco Consulting
MIHAELA VLASEA, Associate Professor, Department of Mechanical and Mechatronics Engineering and Research Co-Director, Multi-Scale Additive Manufacturing Laboratory at the University of Waterloo
SHELLEY FELLOWS, Past-Chair, Automate Canada
STEPHANIE HOLKO Director, Project Development at Next Generation Manufacturing Canada
WALTER GARRISON, Former Advanced Manufacturing Business Consultant for City of Mississauga
Reader Service
Print and digital subscription inquiries or changes, please contact customer service
Angelita Potal
Tel: 416-510-5113
Fax: (416) 510-6875
email: apotal@annexbusinessmedia.com
Mail: 111 Gordon Baker Rd., Suite 400
Toronto, ON M2H 3R1
EDITOR
Jared Dodds jdodds@annexbusinessmedia.com Tel: 416-510-5225
Robot orders in North America increased 4.3 per cent YOY in first half of 2025, A3 data shows
New data from the Association for Advanced Automation (A3) shows robot orders in North America increased by 4.3 per cent and revenue rose 7.5 per cent in the first half of 2025 compared to the first half of 2024.
North American companies ordered 17,635 robots valued at $1.094 billion USD in the first six months of 2025, with automative OEMS leading industry growth with a 34 per cent YOY increase in units ordered.
Other top-performing segments included plastics and rubber (nine per cent) and life sciences/pharma/bio-med (eight per cent).
“The continued growth in robot orders underscores what we’ve been hearing from our members: automation is now central to long-term business strategy,” said Alex Shikany, executive vice president at A3, in a media release.
“It’s not just about efficiency anymore. It’s about building resilience, improving flexibility, and staying competitive in a rapidly changing global market,” Shikany said.
He said if this pattern holds, the North American robotics market could outperform 2024 levels by mid-single digit growth rates by the end of the year.
Cobots accounted for a growing share of the market with 3,085 units ordered in the first half of the year, valued at $114 million.
In Q2, cobots made up 23.7 per cent of all units and 14.7 per cent of revenue, increasingly favoured for their ability to work safely alongside humans and address automation needs in space- or labour-constrained environments.
WORKFORCE
ABB investing more than $130 million in Canadian R&D and manufacturing
ABB has announced an investment of more than $130 million to expand the R&D and production capacity of its advanced power production and grid resilience technologies in Canada.
The investment will combine ABB’s existing Iberville and Saint-Jean-sur-Richelieu facilities in Quebec into a new greenfield location.
This will enable ABB to meet increasing demand in key growth industries, including
utilities, renewables, transportation and residential and infrastructure projects across Canada.
The new 340,000 sq. ft. manufacturing and R&D facility will be more than 33 per cent larger than the two locations it replaces and will incorporate advanced production automation and digital technologies.
Combining the existing ABB operations at Iberville and Saint-Jean-sur-Richelieu will bring more than 600 existing employees from manufacturing, R&D, engineering, testing and support functions into one location.
As new product lines are added and production increases in the future, further jobs will reportedly be created.
“Our investment will support our future growth in Canada, as demand increases with customers focused on grid resilience, power distribution and renewables, and in new transportation, buildings and infrastructure projects,” said Khalid Mandri, president, ABB Installation Products, in a media release.
“The teams in our advanced new facility will deliver products and innovation that will help power businesses and homes across the country. Together with recent investments made in the U.S. and in Europe, the opening of this new facility will be a key part of our global growth strategy,” Mandri said.
The new site is expected to open in mid2027 and will be located in the South Shore region of Montreal.
The new building will integrate clean, energy-efficient electrical equipment and heating systems to reduce energy consumption and cut carbon emissions by over 95 per cent, compared with the two existing facilities.
The project is supported by funding of $16 million from Investissement Québec.
Today, more than 80 per cent of solutions sold in Canada by ABB Installation Products are made or assembled domestically, using 70 per cent locally sourced materials, including 100 per cent Canadian steel and aluminum.
This is ABB’s largest investment to date in Canada. Over the past decade, ABB has invested around US$275 million in its Canadian operations to support advanced
manufacturing, innovation and workforce development.
These investments include upgrades to key facilities, the integration of cutting-edge automation technologies and the expansion of local capabilities and production capacity.
OPERATIONS AND MANAGEMENT
BDC Launches $200M Fund II to Power Canada’s Next Wave of Industrial Innovation
As the global economy undergoes a profound industrial transformation driven by automation and digitalization, BDC is doubling down on Canada’s industrial future with the launch of Industrial Innovation Venture Fund II (I2VF II).
This new $200 million commitment is designed to help Canadian entrepreneurs developing next-generation technologies that will drive productivity and resilience across key industrial sectors.
Building on the success of its first fund launched in 2019, I2VF II will invest in early-stage, high growth companies developing transformative solutions in advanced manufacturing, ag-tech, food-tech, extractive industries and now with an expanded focus on critical minerals, a sector increasingly recognized as vital to Canada’s economic sovereignty and national ambitions.
“Canada’s productivity gap is especially acute in sectors like manufacturing, mining, and agriculture,” said Geneviève Bouthillier, executive vice president, BDC Capital, in a media release. “With Fund II, we’re doubling down our commitment to entrepreneurs building practical, scalable technologies that are highly needed across these industries.”
Fund I addressed a critical funding gap, backing over 20 early-stage companies and demonstrating the strong potential of industrial innovation to boost productivity and global competitiveness.
Notable impacts from Fund I portfolio companies include Acerta Analytics, who stated that it helped a manufacturing facility slash scrap rates by double digits.
“Fund II will target companies that are developing transformative solutions in areas such as robotics, automation, applied AI and industrial software” said Bouthillier in a press statement. “We believe these technologies are key to helping industrial enterprises improve efficiency, reduce waste, and remain competitive in a rapidly evolving global market.”
A VISION FOR AUTOMATION
Learn more about the career of Steve Wardell, director of imaging at ATS Automation and this year’s winner of Manufacturing AUTOMATION’s Lifetime Achievement Award. BY JARED DODDS
PHOTO: JARED DODDS
Research from the Bureau of Labour Statistics, an agency within the U.S. Department of Labour, estimates that the average employee will change careers between three to seven times in their lifetime. This phenomenon is familiar to Steve Wardell, director of imaging at ATS Automation and the recipient of the 2025 Lifetime Achievement Award from Manufacturing AUTOMATION
Just not in the way you’re expecting.
“I’ve had all sorts of opportunities to be part of ATS, but it seemed like every five years it was a different ATS. It was a different company,” Wardell said. “I’ve had four or five different actual careers here at ATS, just in different forms.”
The 36-year veteran of the automation industry has spent his entire career with the Cambridge, Ont.-based automation company, joining after graduating from the University of Waterloo with a degree in math and computer science. The lesson he learned through his education and co-op opportunities was he valued being able to see and hold the end product of whatever he was programming. This passion led him to ATS.
36 years of expertise
Beginning his career on the programming side, Wardell worked with a variety of teams and departments to introduce computer controls into ATS’s custom industrial automation solutions.
“If you want to succeed somewhere, if you want to take it to the next level, you’ve got to do the extra work to get there ... going in with the expectation that ‘I’ll just pick it up as I go’ is an experiment gone wrong.”
When ATS moved to their current location in Cambridge, he was part of the team that founded what amounted to a whole new division for the company, focusing on building high-precision fibre optic device and semiconductor assembly systems. But, as the internet bubble popped, his focus needed to shift. This led to both Wardell’s and the company’s next transition towards automated manufacturing solutions in the life science sector, where he continues to work to this day.
As his career progressed, an opportunity arose in 2010 to assume to a leadership role with a team Wardell was extremely familiar with: the vision team.
“Almost from day one, I was able to work with the team responsible for vision systems. It wasn’t nearly as big as it is now, but we could see the power of it and what it could bring to the overall automation solutions,” he said. “At one point they needed some leadership, they had gone through a lot of different changes. I had been leading the controls side of things for a while and thought I could give
this these guys some of what I’d learned and learn something new in the process.”
This successful transition ultimately led to his promotion to director, where he oversees the entirety of the imaging engineering group, designing and building custom automation solutions for an international client base. As his level of responsibility has grown, so have the requirements for the vision systems incorporated into these solutions.
“Midway through my career, maybe only 50 per cent of the systems we made had any kind of vision on it, and it was maybe only one or two stations,” Wardell said. “Now, every single system going out has a vision station on it and, depending on the size of the system, could have as many as 40 different stations.”
While the presence of a dedicated vision engineering team has become more common across the industry, it was, at one point, unique to ATS, setting them apart from the rest of the market. Wardell carries that legacy forward in his current role, focusing on the creation of distinct vision solutions for a diverse set of client needs.
“I think I’ve learned the most through the failures that have come through my ascension, you learn a lot in a project by going down the wrong path and finding what doesn’t work.”
Perfecting the product
Though ATS focuses primarily on creating tailored, custom automation projects designed to specifically fit a customer’s needs, the necessity for a portfolio of standardized products has increased as the industry has evolved. Wardell has been a part of the design and execution of multiple of these turnkey solutions, some of which have become ATS mainstays.
One was a system designed for an inkjet manufacturer, for whom ATS produced a grit blaster which created holes in pieces of silicon for use in the company’s printer heads. He was responsible for those machines, and what started as a one-off ended up becoming so successful ATS was able to sell upwards of 200 units.
“It was essentially my baby,” Wardell said. “That is the one I associate with when thinking about taking full ownership of a project and seeing it become as successful as it did.”
Additionally, he was integral in the development of a high-speed vision system designed to inspect vaccine vials to ensure compliance with industry requirements. The tool, initially created around 20 years ago, remains in production to this day.
“It was high speed, high precision and technically challenging, but rewarding in the fact that the company that we developed it for partnered with us so they could take it and make it world class,” Wardell said.
Lessons learned
When looking back on his career, Wardell said the greatest learning opportunities are, often, presented under the guise of failure.
“I think I’ve learned the most through the failures that have come through my ascension,” he said. “You learn a lot in a project by going down the wrong path and finding what doesn’t work.”
Wardell noted, however, that learning
from mistakes does not come from passivity, but rather through a recognition of where more effort is required.
“If you want to succeed somewhere, if you want to take it to the next level, you’ve got to do the extra work to get there and not just expect people to give it to you,” he said. “Going in with the expectation that ‘I’ll just pick it up as I go’ is an experiment gone wrong.”
Wardell said he learned this mindset from the great leaders he had the opportunity to work with at ATS, and highlighted Steve Bottoms, one of his first bosses, for setting him on the path that led to where he stands today, and giving him the tools he aims to pass down to his team members.
“I got to see more about what else was out there beyond the day-to-day, and he was always open to pulling me aside and saying, ‘I see what you’re doing here, but if we’re heading down this road we should think about things in a broader perspective,’” he said. “He opened my eyes to think not just laser-focused on the task at hand, but to start to think bigger.”
Moving forward
As Wardell looks back on his career and forward to the future, he sees the continued application of artificial intelligence (AI) into vision solutions as both the ongoing and next major development in the field.
“The machine vision and computer vision world has been the industry leader for the application of AI,” he said. “I don’t think we’re in the breakthrough phase, we’ve gotten through that.
“We’re just learning how to live with it and make the best use of it. We’re going to be seeing, essentially, eyes, on a lot more processes to allow them to be automated more than we ever imagined in the past,” Wardell said.
As for his career, he gave no indication he was slowing down, and had a message for the upcoming professionals hoping to be where he is 36 years into their career.
“When you’re young and you’re just getting into something, you have the ultimate capacity to be flexible,” Wardell said. “Make use of that flexibility. Be open to the next challenge.”
When asked for a summary of his career and his secret to success, his answer was quick, decisive and poignant.
“I’ve been here for 36 years. I’ve never gotten up to go to work and not looked forward to it.” | MA
PHOTO: JARED DODDS
Wardell standing in the ATS Automation factory in Cambridge, Ont., where the company develops custom automation solutions for an international clientele.
ACROSS THE ENTERPRISE
BY JONATHAN GROSS
Your data governance framework
Setting the strategic foundation for operational excellence, unlocking AI, boosting efficiency, ensuring compliance and building resilient digital operations.
At a North American manufacturing plant, a predictive maintenance system flagged anomalies in a key piece of equipment. The alerts should have prompted preventative maintenance. Instead, the maintenance team ignored them.
Why? Because they no longer trusted the system. False positives—driven by inaccurate sensor calibration and inconsistent maintenance logs—had trained the team to dismiss the warnings. Two weeks later, the equipment failed, shutting down the production line for 14 hours. The cost: hundreds of thousands in lost output, expedited shipping and a strained customer relationship.
This is not an outlier. It’s the predictable consequence of poor data governance. When organizations lack trusted, well-managed data, even advanced AI systems produce bad results.
Why data governance matters more than ever
Manufacturers are racing to harness AI, analytics, and automation. The potential gains are significant:
• 25 to 50 per cent reductions in unplanned outages.
• 20 per cent increases in operational efficiency.
• 70 per cent reductions in accounts payable invoice processing costs.
• 90 per cent fewer order-entry errors.
But these benefits depend on one critical factor: trusted data.
The reality is clear: data governance is the prerequisite for AI success.
Are you ready for AI?
Before investing in AI or analytics, manufacturers must assess their data maturity. At Pemeco, we start with six essential questions:
1. Is the data complete?
2. Is it readily available to those who need it?
For many organizations, compliance is the spark for governance.
3. Is it usable—structured, standardized and deduplicated?
4. Is it accurate?
5. Is it secure?
6. Is ownership clearly defined?
Most companies can’t confidently answer “yes” to all six. That’s not failure, it’s a starting point.
One global manufacturer we worked with sought to implement AI-driven inventory optimization. But the source data came from 11 disconnected systems. Units of measure weren’t standardized. Lead times were outdated. Supplier records were duplicated. Without addressing these gaps through governance and cleanup, the initiative would have failed before the pilot.
What is data governance (really)?
Too often, data governance is dismissed as an IT task or compliance checkbox.
In reality, it’s a business-critical framework that underpins digital transformation.
Think of governance as the organizational playbook for data management. It aligns the people, processes, technologies and systems required to ensure data is accurate, secure and usable across the enterprise.
It has three key elements:
• Governance: The rules and decision rights defining how data is managed.
• Ownership: The authority to access, modify, and distribute data.
• Stewardship: The responsibility to maintain accuracy, privacy and compliance.
Critically, governance must bridge strategy and execution, from high-level policy to field-level ERP configurations.
One manufacturer embedded data stewards across departments during an ERP transformation. Their role was not just to approve policies but to enforce them—validating legacy data, establishing rules for creation and maintenance and defining field-level access rights. The result: one of the cleanest ERP go-lives we’ve seen, with immediate improvements in order accuracy and fulfillment speed
Imagine deploying an AI-powered demand planning tool that relies on spreadsheets and siloed systems. Or an accounts payable automation tool pulling from duplicate vendor records. In both cases, bad data sabotages the initiative before it begins. No matter how advanced the technology, it cannot outpace poor data governance. Jonathan Gross is managing director at Pemeco Consulting, a vendor-neutral consulting firm. He leads clients through the entire digital transformation lifecycle, including technology architecture and planning, technology vendor selection, technology procurement, implementation and ongoing optimization.
Governance in action
Because data flows across every function, governance cannot live in silos. It requires collaboration among IT, operations, finance, business leadership and compliance.
Take the example of cleaning up business partner records. It seems technical, but it raises critical governance questions:
• Who owns customer and supplier records—sales, procurement or finance?
• What rules define a duplicate and who approves changes?
• How are workflows structured to validate updates?
• What field-level security prevents unauthorized edits?
Addressing these questions prevents downstream issues like duplicate payments, inaccurate reporting and strained customer relationships. And by embedding compliance and legal stakeholders in the governance team, companies ensure policies also align
with regulations such as GDPR, CCPA, FDA and ITAR. This integration transforms compliance from a bolt-on to a built-in feature of agility and trust.
Governance as a strategic enabler
For many organizations, compliance is the spark for governance. With non-compliance costs averaging $14.82 million per incident, it’s a powerful motivator.
But governance is about far more than risk avoidance. When done well, it becomes a strategic enabler Trusted, accessible data allows manufacturers to:
• Make faster, data-backed decisions.
• Personalize customer experiences in real time.
• Build more resilient supply chains.
• Scale AI and analytics with confidence.
Governance, in other words, is both shield and sword. It protects the organization from risk while unlocking growth opportunities.
Your 90-day data challenge
Building a governance framework may feel daunting, but progress starts with small, deliberate steps. Over the next 90 days, take on this challenge:
1. Assess your data readiness: Identify gaps in completeness, accuracy and availability.
2. Define roles and responsibilities: Appoint data owners and stewards with clear accountability.
3. Establish governance policies: Align them with business processes and ERP configurations.
4. Embed compliance and security: Treat them as design principles, not afterthoughts.
The goal isn’t perfection in 90 days. It’s momentum. Each step builds trust in your data, and trust is the foundation on which AI, analytics and digital transformation thrive. | MA
FAST-TRACKING AGENTIC AI
How companies are achieving faster and more effective AI training through the use of synthetic data.
Manufacturers, get ready. As you may already know, a new level of advanced industrial robotics is coming, sooner than many predicted. And the main driver behind the rapid progress in this area? Using synthetic data to train the AI that runs the robot; the digital robot if you will, generally known as an AI agent or agentic AI.
“Synthetic data is paramount as the robotics industry shifts from pre-programmed robots to generalist, AI-driven robotic systems that operate autonomously in the physical world,” explained Akhil Docca, senior product marketing manager for Omniverse, NVIDIA. “Recent advances in synthetic data have revolutionized robot training by allowing robots to learn a diverse set of tasks, virtually.”
BY TREENA HEIN
To understand why synthetic data is such a game-changer, it’s important to review what happens when an AI agent is trained with real-world data. This entails setting up one or more robots in a lab space that mimics the corresponding industrial setting, a process that is typically difficult and expensive. It’s also time-intensive and painstaking, with a data collection rate limited to a robot’s operation speed.
years, said Roberge, to create simulators and virtual training environments for AI agents. Now there are many of these platforms, with a handful leading the pack. Roberge explained that with advanced simulators and full digital twins of factories or warehouses, “engineers now create high-fidelity artificial datasets that replicate sensor streams
Sanctuary AI has used their proprietary Reinforcement Learning approach to enable in-hand reorientation under an extreme disturbance, including a 500 G load that was not encountered during training.
“Improved simulation allows the potential of reinforcement learning to be applied to complex platforms... providing an opportunity to train complex manipulation behaviours in the real world.”
— Olivia Norton, CTPO, Sanctuary AI
(e.g. RGB images, depth, force or torque, even tactile feedback) across a vast range of scenarios.” Simulators can even output detailed synthetic tactile signals, he said, which enable AI agents to master delicate object
manipulation, an area in which Roberge and his grad students have recently published two papers.
Of course, the quality of synthetic data is improving all the time, closing what’s called
the ‘sim-2-real’ gap. Because of its increasing accuracy and broad usefulness, the use of SDG is predicted to surpass real data in AI models by 2030, according to tech industry consulting firm Gartner.
This gap, according to Olivia Norton, cofounder and CTPO of Sanctuary AI, is broadly being narrowed right now by advances in high-fidelity simulation, neural rendering and photorealistic domain randomization. “ To fully close the sim-2-real gap,” she said, “we still need to see improvements in contact dynamics, actuation dynamics, deformable material simulation, touch sensing simulation, high-fidelity collision geometry representation, multi-camera simulation and simulation speed.”
Currently, Norton reported that it’s becoming easier and more effective to train humanoid systems (including robot hands) in increasingly complex scenarios. “Improved simulation allows the potential of reinforcement learning to be applied to complex platforms such as dexterous high-performance hands, providing an opportunity to train complex manipulation behaviours in the real world,” she said, “and also inform design iterations of the hand.”
In this area, Sanctuary has recently achieved a unique feat. “Our proprietary Reinforcement Learning approach has enabled in-hand reorientation under an extreme disturbance – a 500 G load that was not encountered during training,” Norton said. “This is first of its kind and a significant milestone.”
Enjoy working with your hands
Where SDG will be felt the most Roberge believes that robotic applications that demand dexterity and adaptability are where SDG will have its deepest impact in manufacturing.
“Sectors are transitioning from static production lines to highly diversified, unstructured environments where robots must handle constant variation,” he said. “I collaborate with numerous industrial partners in the manufacturing sector who face this reality every day, [who need] more agile and versatile robots capable of performing far more complex tasks than the limited, repetitive operations they once handled.”
What’s ahead
Beyond advances in simulators, Docca pointed to advances in generative AI as important in further narrowing the gap between virtual training and real-world performance.
“This will enable manufacturers and logistics leaders to reduce deployment times and increase automation across a wide range of tasks,” he said. “Additional use cases could involve the use of AI agents to monitor and track certain manufacturing processes like cable assembly for electronics chassis.”
Roberge added that simulations with SDG are increasingly being used with complex robotic assembly tasks that are rich in contact mechanics, such as insertion and fastening. In quality control, he said synthetic data is now enabling engineers to generate thousands of examples of rare defects directly from CAD models.
In Roberge’s view, there are three factors that will play a major role in making closing the sim-2-real gap. One is the rise of versatile ‘foundation models’ for robotics, analogous to GPT for language. These models will be pretrained using billions of simulated interactions and all customers will have to do is fine-tune them for specific robots and tasks.
Another is the democratization of advanced robotics. “As powerful simulation platforms become more accessible and cloudbased, small and medium-sized manufacturers will be able to develop, test and deploy highly-sophisticated, adaptable robotic solutions at a fraction of the traditional cost,” Roberge said, “leveling the playing field and accelerating innovation across the sector.”
The third factor is the increased ease of creation of synthetic data. “Generative AI models will produce photorealistic, physically-accurate training scenarios from simple text prompts,” Roberge explained. “This will
create a loop in which increasingly-capable AI builds better simulations to train even more powerful AI, an approach akin to metalearning.”
However, Roberge does not expect actual AI training speed to increase much for a while, maybe a year or more.
He also pointed to the reliability of the
data as a potential bottleneck. “You can now generate a lot of data fast, but accuracy is still a concern,” he said. “Sometimes you can rely completely on synthetic data. It will be enough in some cases. But in others, you need real-world data. And it takes time to acquire that. It’s still a long and tedious process.” | MA
Modular Safety Relay:
WHAT IS CANADA’S LEVEL OF ROBOTICS READINESS?
How the federal government can better support robotics implementation in the manufacturing sector through a national robotics strategy.
BY JARED DODDS
The manufacturing industry has been grappling with a lack of skilled labour for some time. A recent push from policymakers to address this gap through investment in training is merely the first attempt to identify a solution for a problem many businesses have faced since their inception. The realities of the skilled labour gap and productivity crisis are clear: there is not enough personnel to fill these crucial roles.
Where the human solution is lacking, however, the robotic solution can succeed. One proven method of addressing the lack of skilled labour is introducing automation into the manufacturing process, eliminating the need for humans to complete the repetitive, mundane or dangerous tasks necessary to keep the business running and freeing up their time to address more complex problems.
This method has proven successful in countries including Germany, Japan, and Sweden, all of whom have a manufacturing sector that accounts for approximately 28
per cent of GDP, very similar to Canada. In Germany, approximately 84 per cent of manufacturing companies have integrated some form of robotics or automation into their operations, reported the International Trade Administration. In Japan, that number is 80 per cent and in Sweden somewhere between 70 to 80 per cent.
In Canada? 42 per cent.
Where’s the problem?
Despite having a robust manufacturing sector and a surplus of home-grown innovators in the robotics sector, we are lagging behind. What separates Canada from these other countries is not the capability to integrate these solutions, but rather the support from the government required to do so.
“If I had to give a grade right now for robotics readiness, I’d give it a C+. But I want to be really clear that it’s not about a lack of talent or potential or technological ability,” said Hallie Siegel, chief executive officer of the Canadian Robotics Council.
“I see Canada as the student in the class that has a lot of natural inherent ability, but who’s underperforming. Because the system around them isn’t supporting them properly, it’s not giving them the right tools or attention of encouragement.”
— Hallie Siegel, CEO, Canadian Robotics Council
“I see Canada as the student in the class that has a lot of natural inherent ability, but who’s underperforming. Because the system around them isn’t supporting them properly, it’s not giving them the right tools or attention and encouragement,” Siegel said.
A major difference between the majority of global leaders in robotics and Canada is a coordinated national strategy outlined by the federal government to support manufacturing organizations throughout the adoption process.
Germany has the Robotics Research Action Plan and High-Tech Strategy 2025, which both focus on enhancing robotics research and development, Japan has the New Robot Strategy as part of their larger Society 5.0 concept, the United Kingdom recently introduced their Smart Machines Strategy 2035 with the goal of transforming into a global leader in robotics and the United States has the National Strategy for U.S. Leadership in Advanced Manufacturing.
In contrast, Canada employs a decentralized ecosystem of programs, clusters and other agencies to promote robotics adoption in the manufacturing sector. Some of these players include Next Generation Manufacturing Canada, the Strategic Innovation Fund, Innovation, Science and Economic Development Canada and more.
The question, then, is does a singular voice and message from the top really move the needle? Søren Elmer Kristensen, chief executive officer of Odense Robotics, Demark’s robot
and drone cluster, said his country’s national strategy, launched in 2020, has been integral in sending a message to the private sector that the government is serious about robotics adoption and will support these businesses as they embark on their integration journey, building the trust and confidence manufacturing leadership needs to take the leap.
“Leadership from the top, strong political will and strategic focus from the government are essential to align regulation, funding, education and industry initiatives, making sure all parts of the ecosystem pull in the same direction,” Kristensen said.
That level of trust, as demonstrated by our rate of robotics adoption, is severely lacking in Canada on a national level. Siegel highlighted some cities including Mississauga, Brampton, Waterloo and Edmonton who are introducing industrial strategies, but agreed the gap at the federal level was holding back Canada’s manufacturing sector.
“Canada ranks seventh in the world for service robotics manufacturers and fifth on a per capita basis. We punch above our weight in robotics talent and exports, but misaligned policy and investment is holding us back from putting that expertise to work at home.”
Recent developments
With the election of Mark Carney as prime minister in April, there has been some roundabout change in the governments approach to
robotics, namely through the introduction of Evan Solomon as the minister of artificial intelligence and digital innovation in May 2025.
This emphasis on AI will, in theory, lead directly to further investment in the hardware application of AI solutions, namely, robotics. But Siegel would like to see that investment be more defined.
“We need to empower the new AI ministry with a clear mandate for physical AI. Unlike software-only AI, adopting robotics has higher upfront costs and needs more patient capital,” she said.
“Allocating a portion of existing AI funding towards robotics, and making it an explicit national priority, just as our government has done for quantum computing and clean tech, would send a strong signal to manufacturers that they’ll get the right supports to adopt and innovate,” Siegel said. “The irony is that robotics offers a quicker path to return on investment than quantum or cleantech, making it a better value for taxpayers.”
Kristensen noted that when the investment is defined in the public sector, it creates more engagement in the private sector, leading to further funding and the potential for innovation.
Ideal outcome
If the government decided to move forward with a national robotics strategy, what would it look like? For Siegel, a big part of what this framework needs to accomplish is an
updated understanding of robotics, recognizing its capability to enable cross-sectoral growth not in the future, but right now.
“Our current national policy still classifies robotics as an emerging technology, but that framework is from 2019, and there are plenty of commercial solutions ready right now,” she said.
“The government needs a more nuanced definition of robotics if it’s going to effectively support industry to adopt and innovate in this space,” Siegel said.
That redefinition would come with the promotion of advanced manufacturing, highlighting the exciting industry opportunities to encourage younger talent to move into and stay in the field. It would also require an investment in workforce training programs to upskill personnel and prepare them to work alongside robotics, stemming the fear of job loss and encouraging technology adoption.
Developing that trust is key, reported Kristensen, who said the level of acceptance of robotics in Denmark is as high as it is because personnel had the opportunity to work alongside robotics and see the possibilities provided.
“When we’re looking at the manufacturing floor, we see a high level of trust because we have had quite a long tradition of robotics in Odense,” he said, referencing the introduction of robotics to the city’s shipyard over 40 years ago. “We also have the support from labour unions who see they’re [robots] are there to support their work and not replace them.”
Future outlook
Prime Minister Carney’s government will table its first budget in October, so there is still a possibility a robotics strategy is on the way. Siegel highlighted two possible paths the government could move down, with only one leading to a positive outcome.
“I see two scenarios. One of them is business as usual, carrying on thinking that the same policies and strategies that we’ve had are going to achieve different results, and they won’t,” she said.
“The second scenario is that we act now to get Canadian manufacturers the supports they need to automate. While many worry that we’re behind, I use that C+ as a warning to get people moving.
“All the ingredients are here for Canada to be a global leader in this space, and it will create resilience across the whole manufacturing ecosystem and beyond, Siegel said. “We can do these things. I totally believe it.” | MA
AUTOMATION IN A RAPIDLY CHANGING WORLD
As new technology expands the boundaries of automation, the task of controlling automated processes is becoming increasingly complex. AI-powered adaptive control systems that learn from their environment are becoming essential in many scenarios.
BY JACOB STOLLER
The attempt by General Motors during the 1980s to create a fully automated “lights out” factory is possibly the most expensive, and talked about, automation failure in history. What killed the project was not that the thousands of robots deployed couldn’t build cars – it was that the entire ecosystem of robots and control systems wasn’t flexible enough to keep pace with the rapid changes demanded by the automotive market.
Since then, the flexibility challenge has only gotten tougher as manufacturers adapt to high-mix, low volume demand. According to research by MIT’s Industrial Performance Center, productivity gains from automation have been often cancelled out by high changeover costs and other expenses required to adjust to changing circumstances.
“Our research shows that what a company gains from automation in productivity it tends to lose in process flexibility,” wrote MIT
professors Ben Armstrong and Julie Shah in the Harvard Business Review. “And when firms do adopt automation technology, studies show, they end up hiring more workers, not fewer, as they become more productive.”
As Francois Giguere, Chief Technology Officer for Montreal-based automation manufacturer Vention explained, there are many scenarios in which it is very challenging to get an acceptable ROI from automation, even in large companies.
“If you go on factory floors of big Fortune 500 companies, you’ll see incredibly complex and sophisticated automated machines,” said Giguere, “and right beside, you’ll see applications that are completely not automated. These are processes that they just threw bodies at because they are high-mix low-volume, and the ROI from automation is not straightforward.”
For example, a beverage plant might have highly sophisticated automation for filling and capping thousands of bottles per hour, but struggle with automation at stages such as machine tending or palletizing.
“These cells get reconfigured all the time, so that’s where you have the need for flexibility,” said Giguere. “In the past, people just threw bodies at these processes. But now, companies want to get savings, so they are targeting these unautomated areas. So, what we are doing is utilizing AI and vision
systems to make the machines more flexible in terms of infeeds and outfeeds.”
The combination of AI and vision systems allows a robot to act not just according to fixed programs and algorithms, but according to what is presented to the robot in real time, the same way a human would. This is far more complex than most people realize.
“Just grabbing a coffee cup is super complex,” said Giguere. “There are ways in which you can grasp and hold it, and ways you can’t. Figuring that out is a very difficult problem for a robot.”
Vention breaks down the problem of grasping and moving a part into three categories. The first is pose estimation, where
the system determines the exact position and orientation of part. The system then determines the optimum way to grasp the part. Finally, a 3D motion analysis determines how the robot will move the part.
The technology leveraged to accomplish this includes a wide range of cutting-edge technology, including 3D modelling based on synthetic data, GPU-powered parallel computing, vision systems that can adapt to varied lighting conditions, and many components of AI.
Extending the boundaries
Adaptive control isn’t just about imitating humans – it’s also about extending the boundaries of what we can control.
“Adaptive control can operate in situations where reality is changing faster than what a classical controller can typically handle,” said Sheldon Fernandez, AI strategist and founder and former chief executive officer of DarwinAI.
A key challenge for classical control systems is non-linearity, that is, situations where a small fluctuation can lead to an amplified effect on an overall system. “Classical control systems don’t do a great job of handling non-linearity,” said Fernandez. “As an example, a client of ours in the aerospace sector found that the latest form of neural network-powered AI in adaptive controllers handled those nonlinear situations a lot better.”
Autonomous drones are a key area where this technology will pay off. “This small fluctuation in one of the variables could amplify
the trajectory if you don’t model it properly,” said Fernandez. “So that’s an example of why adaptive controls are so powerful.”
Wind turbines and precision mining are other areas where adaptive control systems are being deployed to prevent small variations from triggering large deviations from the desired outcome.
On manufacturing lines, the same approach can also be used to prevent small deviations from escalating into major quality headaches. As Fernandez explained, manufacturing machinery is now equipped with numerous IoT devices that are collecting data that can be useful in the adaptive control process.
At DarwinAI, for example, the company aspired to take data from the end of the line and feed it back to controllers earlier in the process in order to influence the outcome from that earlier point in the process. “The aspiration is that we can do intelligent data collection,” said Fernandez, “and that can be a feed into the beginning of the system so that it changes things, perhaps to a very small degree, to increase the quality at the end of the line.”
Bringing adaptive control into a manufacturing environment will introduce technological challenges that are new to many manufacturers. One will be adapting legacy equipment and control environments. Another is upgrading skillsets so that people understand not so much how AI works, but what it does with data, and the kinds of predictions it is able to make.
There will be regulatory issues as well in some industries, as AI-powered automation can be opaque and make it difficult to verify compliance.
There’s also a danger of rushing to adaptive controls when there may be better options. “Obviously all the technologies we’re discussing will continue to advance, becoming smarter, more adaptive, easier to use, and less expensive,” said James Taylor, chief commercial officer of OnRobot. “The buzz phrase of the hour, artificial intelligence, will certainly play a part there, and that research is absolutely important. But it’s easy to get caught up in chasing automation holy grails when the reality is that the industry could make huge strides if manufacturers adopted and benefited from the smart technologies that are already available and proven.”
Perhaps most important, companies will have to develop a business culture that fosters leading edge automation technology not just as a means for replacing people, but for improving the work that the company does.
“A lot of this is imagining the world that we are not imagining,” said Fernandez. “Maybe twenty years ago, somebody might have looked at the internet and said, ‘how can I use this to make taxis more efficient?’ And there were ways you could do that. But it took somebody else with a very different mindset to connect those, and in doing that they completely reinvented how we think about shared rides today. And so, that same kind of creative thinking needs to be done in various industries.” | MA
Airflex Failsafe Caliper Brake
Danfoss Power Solutions has launched its Airflex ACB-F3 Failsafe Caliper Brake. The ACB-F3 features a clamping force of 20 to 120 kN and a braking force of 16 to 96 kN. Release oil pressure ranges from 40 to 190 bar (580 to 2,756 psi). The brake features a compact design that provides reliable, stable braking torque, with its symmetrical opposed piston caliper design suited for hightorque, high-energy applications. The brake is available with organic or sintered friction pads. An anti-corrosion version is available for use in extreme environments.
danfoss.com
MOTION CONTROL
Nano High Power Digital Servo Drive
Copley Controls has announced the release of its Nano High Power (Nano HP) and Nano Plus High Power digital servo drives for brush and brushless motors. Both series expand Copley’s compact Nano family and feature redesigned power boards for higher current
output, making them ideal for traction motor applications in autonomous mobile robots, automated guided vehicles and other space-limited mobile robotics requiring higher-power embedded modules. Both drives operate from 9 to 90V DC input voltage and deliver up to 140A of continuous current and support EtherCAT and CANopen communication protocols for real-time data exchange, with the Plus-HP supporting additional networking protocol options.
copleycontrols.com
LOGISTICS AND MATERIAL HANDLING
Rapid Series Palletizer
Vention has released its nextgeneration Rapid Series Palletizer, which, while available now, will make its North American debut at Pack Expo Las Vegas. Built on Vention’s FastFactory plug-and-play hardware ecosystem, key features include a vertical reach up to 136 inches with an optional riser, a redesigned base supporting both pallet jacks and forklifts for easy relocation, access to live video assistance from automation specialists and built-in analytics to monitor operations remotely and unlock actionable insights. When combined with Vention’s modular conveyors, box erectors and case packers, the Rapid Series Palletizer is part of an interconnected end-of-line ecosystem, all managed through a unified software interface. vention.io
SENSORS AND VISION
RVD Series Hall Effect Angle Sensors
NewTek Sensor Solutions has launched the RVD Series Hall Effect Angle Sensors designed to meet the demands of industrial environments. Utilizing Hall Effect technology, these sensors reportedly provide rotational position (angle) measurement, offering high performance and dependability in the harshest conditions. NewTek RVD Hall Effect Rotary Position Sensors operate over a wide temperature range of -30°C to 80°C, exhibiting excellent linearity of +0.3 per cent of full range output. The non-contact design of the sensors measures rotational position without mechanical moving parts, reducing the wear and extending the sensor’s operational life. newteksensors.com
ENERGY MANAGEMENT AND ENCLOSURES
Pro 2 Series Power Supply with Redundancy
WAGO has released the Pro 2 Series Power Supply with Redundancy, offering a fully redundant power supply system without the need for separate equipment due to integrated MOSFET functionality. The design of the redundancy offering shares many of the same features as the original Pro 2 family, including communication modules support for I/O Link, Modbus RTU Modbus TCP, EtherNet/ IP+MQTT protocols and smart load management with TopBoost and PowerBoost. The new model offers a n + 1 configurations, the load is designed to be powered by the n power supplies. wago.com
INDUSTRY WATCH
BY PAUL HOGENDOORN
Paul Hogendoorn is a serial entrepreneur that has built multiple manufacturing and manufacturing technology companies. He has been a regular contributing columnist for 15 years. For more insight or information from Paul on related topics, please contact paul@tpi-3.ca or visit TPI-3.ca
Tariff talk: Time to place your bets
For contextual purposes, let me put these three biases out there: the first is that I am a huge “America” fan; the second is I believe manufacturing created (and still sustains) the middle class; and the third is, I am in agreement when airline flight attendants instruct you to put your own oxygen mask on before helping your children or others put their mask on.
Regarding my first bias, this is not to say that I am a fan of one party, or one period, or in agreement with all policies or leaders’ positions. But if I look at “America” in the context of 1776 to today, and more specifically from the 1940’s to the current day, my belief is that most people who are living in what’s described as “western democracy” are benefactors of America and its history-altering constitution.
Never before this period of human history have so many average citizens had the opportunity to live in a house they own (mortgage or not), have discretionary income and have recreational time to spend at their own discretion. Go back even 150 years, and it’s only the nobles, blue bloods and elites that had these three things.
My second bias comes from the net economic benefits the manufacturing revolution brought not just to countries and companies, but to families, individual workers and communities. The value of the output from these primary industries (mining, farming, forestry and fishing) gets multiplied many times over by secondary industries (manufacturing), and much of that multiplied economic benefit gets distributed directly over a far broader base of people.
And my third bias is simple: the best way to be able to safeguard the health and wellbeing of others is to make sure you are healthy and capable yourself. Unselfishness, charity and generosity are all virtuous but not being a good steward of what’s been entrusted to you is irresponsible.
We need more manufacturers (ie. real GDP generators, real employers) speaking up and getting involved, and less politicians aiming to curry favour from their base.
It’s with those three biases in mind that I take a longer and more considered view of the current tariff policies of Canada, the U.S. and all the countries now affected by those decisions. To me, it comes down to this: a healthy and prosperous America has historically always been a good economic partner and foundation to build on and with for countries that share similar values. And manufacturing has historically been a key driver for their whole economy, specifically for their blue collar and middle-class sectors.
It might seem selfish to many right now for U.S. policy to favour U.S. manufacturers or encourage more companies to try to manufacture in America, but longer term, it leaves countries like Canada with important decisions to make: do they bet on America in the long term and position themselves to be preferred partners with them going forward, or do they bet against them and put policies in place that distance them even further?
I know which course I would choose. We, or the collective western society, need America to succeed and continue to be the primary economic driver for western democracy, and we need manufacturing to continue
to be the primary economic engine for many in our middleclass. Manufacturers are salt-of-the-earth people. They build genuine win-win relationships that are measured in decades, even generations, not in years or terms. Right now, we hear feisty slogans and harmful political rhetoric, aimed at encouraging voters to believe there are better markets and customers for our industry to aim for, gambling with the long-term future of our industry for their short-term political gain. We need more manufacturers (ie. real GDP generators, real employers) speaking up and getting involved, and less politicians aiming to curry favour from their base. Manufacturers know how long it takes to build a sustainable business and earn the trust of good customers – because they do it day in and day out, decade by decade, generation by generation. They know what’s at risk. If it was up to me, I would take the long view – historic and future. I’d place my bets on America succeeding, allowing our manufacturing sector to succeed as they achieve their success. They may have to succeed first (because they are the leaders and the first to act), but in truth, it’s always been that way.
That’s where I’d place my bet. | MA
ROBOTS IN MOTION
SIX STEPS TO ACHIEVING ACCURATE ROBOT POSITIONING
In the world of robotics, knowing the exact position of your robot in the workstations is crucial. Even minor deviations can significantly impact performance, whether it’s due to imprecise docking of mobile robots, inaccurate part carrier feeding, or deviations occurring in gripper/tool positions. Enter the PLOC2D SmartMark – the solution to all your positioning challenges.
