Hit enter to search or ESC to close
Trending Now
MR. ROBOT Sdn. Bhd. and TROOPERS Innovation Sdn. Bhd. are exploring a new approach to training humanoid robots, one that draws directly from how people perform everyday tasks in real workplaces.
The two companies have signed a Memorandum of Understanding (MoU) to explore the collection and use of real-world training data for humanoid robotics and embodied AI, technology that enables robots to understand and perform tasks in the physical world.
The two companies have signed a Memorandum of Understanding (MoU) to explore the collection and use of real-world training data for humanoid robotics and embodied AI, technology that enables robots to understand and perform tasks in the physical world.
 |
| From left: Joshua Tan, CEO and Co-Founder of TROOPERS and Edwin Yap, Founder and CEO of MR. ROBOT |
At the centre of the collaboration is TROOPERS' workforce network, which includes more than 500,000 registered workers, over 270,000 hours of frontline work each month and more than 1,000 enterprise clients. This will be paired with MR. ROBOT's expertise in robot training, AI model development and data delivery.
By bringing the two capabilities together, MR. ROBOT and TROOPERS will explore how data generated by people performing tasks in actual workplaces can complement data generated by robots. The aim is to support the training of humanoid robots for tasks in commercial environments.
Using Real Workplace Tasks to Train Humanoid Robots
TROOPERS' network covers frontline industries such as retail, warehousing, food and beverage, and events. These environments provide access to a variety of everyday tasks, ranging from stocking shelves and picking and packing products to preparing food and beverages.
Under the proposed collaboration, TROOPERS will coordinate participating workplaces, permissions and worker consent, while also facilitating first-person recordings of these tasks. Captured from the workers' perspective, the recordings will provide information on the movements, actions and workflows involved in completing everyday work.
The human-generated data will complement MR. ROBOT's existing robotics data operations in Malaysia. Its purpose-built Data Collection Centre currently houses 20 AgiBot G2 humanoid robots operating around the clock, while TROOPERS will provide a pathway for structured first-person task data to be collected directly from real workplaces.
Through its Data-as-a-Service (DaaS) offering, MR. ROBOT will handle data collection, annotation, quality review and dataset preparation. This brings together two different but complementary sources of information: human-generated egocentric data showing how people perform tasks, and robot-generated data capturing robotic movement and behaviour.
By bringing the two capabilities together, MR. ROBOT and TROOPERS will explore how data generated by people performing tasks in actual workplaces can complement data generated by robots. The aim is to support the training of humanoid robots for tasks in commercial environments.
Using Real Workplace Tasks to Train Humanoid Robots
TROOPERS' network covers frontline industries such as retail, warehousing, food and beverage, and events. These environments provide access to a variety of everyday tasks, ranging from stocking shelves and picking and packing products to preparing food and beverages.
Under the proposed collaboration, TROOPERS will coordinate participating workplaces, permissions and worker consent, while also facilitating first-person recordings of these tasks. Captured from the workers' perspective, the recordings will provide information on the movements, actions and workflows involved in completing everyday work.
The human-generated data will complement MR. ROBOT's existing robotics data operations in Malaysia. Its purpose-built Data Collection Centre currently houses 20 AgiBot G2 humanoid robots operating around the clock, while TROOPERS will provide a pathway for structured first-person task data to be collected directly from real workplaces.
Through its Data-as-a-Service (DaaS) offering, MR. ROBOT will handle data collection, annotation, quality review and dataset preparation. This brings together two different but complementary sources of information: human-generated egocentric data showing how people perform tasks, and robot-generated data capturing robotic movement and behaviour.
Turning Everyday Workflows into Robotics Data
Rather than relying solely on controlled robotics environments, the collaboration will examine how work is actually carried out in commercial settings. Data will be collected and analysed from tasks including stocking, picking and packing, and food and beverage preparation.
The resulting insights will complement robot-generated datasets, supporting the development and refinement of AI models for humanoid robots.
Edwin Yap, Founder and CEO of MR. ROBOT, said: “Real-world task data gives us insight into how work is actually performed beyond a controlled robotics environment. By combining human-generated data with robot-generated data, we can explore new ways to develop AI models that help humanoid robots operate more effectively in commercial settings.”
For TROOPERS, the collaboration also expands on its existing workforce infrastructure by exploring how its relationships with frontline workers and commercial workplaces can contribute to robotics data collection. At the same time, the proposed approach will maintain appropriate standards around workplace access, worker participation and data governance.
Joshua Tan, CEO and Co-Founder of TROOPERS, said: “Frontline workers understand the realities of how work gets done every day. Through this collaboration, TROOPERS can help bring that practical experience into the development of robotics while ensuring worker participation, workplace access, and data governance are managed responsibly. We see this as an opportunity for the workforce to play an active role in shaping how automation evolves across the region.”
Rather than relying solely on controlled robotics environments, the collaboration will examine how work is actually carried out in commercial settings. Data will be collected and analysed from tasks including stocking, picking and packing, and food and beverage preparation.
The resulting insights will complement robot-generated datasets, supporting the development and refinement of AI models for humanoid robots.
Edwin Yap, Founder and CEO of MR. ROBOT, said: “Real-world task data gives us insight into how work is actually performed beyond a controlled robotics environment. By combining human-generated data with robot-generated data, we can explore new ways to develop AI models that help humanoid robots operate more effectively in commercial settings.”
For TROOPERS, the collaboration also expands on its existing workforce infrastructure by exploring how its relationships with frontline workers and commercial workplaces can contribute to robotics data collection. At the same time, the proposed approach will maintain appropriate standards around workplace access, worker participation and data governance.
Joshua Tan, CEO and Co-Founder of TROOPERS, said: “Frontline workers understand the realities of how work gets done every day. Through this collaboration, TROOPERS can help bring that practical experience into the development of robotics while ensuring worker participation, workplace access, and data governance are managed responsibly. We see this as an opportunity for the workforce to play an active role in shaping how automation evolves across the region.”
