Extend Robotics accelerates robotic learning for real-world industrial automation, in collaboration with NVIDIA and Leyland Trucks
- louisegoldsmith1
- Oct 2
- 4 min read
September 2025, Extend Robotics, in collaboration with Leyland Trucks, accelerates robotic learning for real-world industrial automation with NVIDIA technology.
In modern vehicle manufacturing, robotics are becoming a central driver of safer, faster, and more efficient production. Spray painting vehicles, managing high-voltage electrical assemblies, carrying out precision inspections, and performing repetitive maintenance are all essential tasks, but each carries significant risk for human workers. Robots are uniquely suited to step into these environments, reducing hazards while maintaining high levels of accuracy and output.
A global survey conducted by QNX, drawing on insights from over 1,000 industry decision-makers, found that 65 percent of automotive businesses are already using robotic systems in their operations. Confidence in safety is strong, with 82% of automotive respondents saying they trust robots to carry out essential tasks, including 24% who say robots are “fully” trusted to perform these tasks “consistently and autonomously.”
The challenge lies not in recognising the value of automation, but in achieving integration at speed and scale. Traditional approaches to robotic training require lengthy periods of programming and constant operator involvement. Teaching robots to manage delicate assembly tasks, conduct precision welding, or carry out micron-level inspections can take weeks or even months, consuming both time and resources.
Extend Robotics has developed a solution that moves beyond these limitations. By combining immersive extended reality teleoperation with advanced robot training systems, the company accelerates the pace of robotic learning across a wide range of hardware platforms. Starting with the NVIDIA Isaac-Mimic Blueprint, human experts can demonstrate tasks in XR, automatically generating rich datasets and virtual environments for AI model training. The collected data is then used to teach robots how to complete complex jobs inside a safe, realistic computer simulation, in NVIDIA Isaac Lab, before those robots ever set foot on a factory floor for faster learning capability, reduced risk, and no downtime in production. Finally the learning skills can be deployed on NVIDIA Jetson AGX Orin for real world execution.
And these benefits are already being demonstrated in real-world environments.
In a factory near Preston where around sixty trucks roll down the assembly line each day, a changing mindset is unfolding. Extend Robotics, a London-based tech innovator, has partnered with Leyland Trucks to explore a bold vision: a world where human intuition meets robotic precision, safely, swiftly, and at scale.
The heart of the initiative lies not in replacing humans, but in empowering them. For two days, Extend Robotics orchestrated a feasibility trial inside Leyland’s state-of-the-art facility, immersing human operators in a virtual realm where they could guide humanoid robots through difficult and hazardous tasks.
Among the most striking scenes was the robot’s precise installation of a Master Service Disconnect. The task, notoriously dangerous for humans (and usually done under thick insulation and protective gear), was successfully executed by a robot after just two days of training. By relying on simulation for validation, training time and costs were reduced by more than 60 percent while ensuring reliability from the outset.
At Leyland, the implications are tangible. The robots completed tasks remotely in just moments, reducing risks to human workers and avoiding the longer timescales such jobs would normally require. Operators who lacked any VR experience reported feeling immediately comfortable in the virtual cockpit. "It’s like your hand is that hand," one participant said, capturing that rare moment when technology and humans partner.
Witnessing this offered a glimpse not only of what’s possible, but what’s inevitable when technology serves expertise rather than replaces it.
Meanwhile, the broader industrial world is watching. Safety hazards and labour constraints are pushing manufacturers to rethink automation, not merely as a cost driver but as a force for human-centric productivity. The feasibility trial at Leyland is the first chapter of a larger narrative where XR interfaces capture expert knowledge and turn it into scalable training for robots. Beyond heavy vehicle assembly, the applications are expanding. In agriculture, for example, VR-driven teleoperation is training fleets of robots to remotely pick strawberries, scaling up to 500 seasonal workers virtually overnight.
Behind the VR headsets and simulated environments lies a more compelling reality: this is industrial storytelling in action. Experts are no longer confined to physical constraints but can teach robots from anywhere. Workers no longer need to risk exposure to dangerous environments. Factory floors no longer need to pause for retraining cycles that once dragged on for weeks or months. Human insight flows directly into AI, and robotics arrive day one ready to perform.
Chang Liu, CEO and Founder of Extend Robotics explains “Extend Robotics’ work with Leyland signals more than just another tech demo. It marks the beginning of a future where factories are places of augmented excellence, where human knowledge travels through virtual hands to robotic bodies, delivering speed, precision, and safety in equal measure. As more sectors, from EV battery assembly to harvest fields, get involved in this revolution, the question is not whether robots will assist in production, but how human ingenuity will guide them there.”
Learn how you can unlock the potential of autonomous and remotely operated robots, making workplaces safer, smarter, and more scalable by requesting a demo of our software here.
