Driving Motion Precision: How Physical AI is Converting Data into Action
Autonomous manufacturing is entering a new phase where robotics, Physical AI, and semiconductor innovation converge to transform how factories sense, decide, move, and adapt. In a detailed conversation with Anwesh Koley of ELE Times, Allan Lagasca, Director, Smart Industrials Segment, Robotics Segment Strategic Program, STMicroelectronics, delved into how robots have evolved into intelligent physical systems combining perception, edge AI, motion control, actuation, power management, connectivity, and functional safety.
ELE Times: Share your thoughts on physical AI and how you see it transforming machine operation?
Allan Lagasca: Physical AI is the bridge that converts data into action. It brings together AI, sensing and control so machines can act in the physical world. The challenge is evident in a robotic hand that resembles a human hand. With six motors controlling the robot and 21 data points generated by the hand, AI, sensing, processing and control must work together precisely, even as the hand moves constantly. This is what the machines do when they sense, reason and respond in real-time, rather than executing a predetermined sequence of instructions.
ELE Times: What is the vision of STMicroelectronics towards Smart Industrials?
Allan Lagasca: Our vision for Smart Industrials is to advance from individual products to modules and system-level solutions. Our Robotics Segment Strategy integrates sensing, edge AI, processing, motor control, power, connectivity and safety to support robotic hands, joints and cameras within a coherent architecture. Through reference designs and collaboration with module and ecosystem partners, we help customers simplify integration, improve motion precision and energy efficiency, reduce system cost and accelerate deployment. This positions ST as a system-level technology partner, addressing customer applications rather than simply supplying components.
ELE Times: How is STMicroelectronics translating AI language into practical robotics applications for the Indian market?
Allan Lagasca: We translate AI into practical robotics through development platforms and ecosystem collaboration. Our compact 6DoF Robotic Hand with AI Gesture Recognition Control demonstrator, powered by the STM32N6 MCU, combines motion sensing and processing. For India, the opportunity is to strengthen the robotics supply chain around precision joints, hands and modules, while developing datasets for robotic AI training and building on emerging prosthetics capabilities. We aim to connect Indian start-ups, new robotics players and manufacturers with the global robotics ecosystem, helping them develop practical applications and participate in global supply chains.
ELE Times: How does automotive-grade electronics apply to robotics? How do you ensure safety, real-time control, harsh environments, and high uptime?
Allan Lagasca: Both cars and robots need to operate reliably around people. Proven quality, long service life, endurance and suitability for challenging environments are all hallmarks of automotive-grade electronics. Because humanoids are also mobile and interact directly with humans, safety becomes an even more important consideration. Safety needs to be designed into the system because even if a robot meets a particular standard, a sensor failure or an incorrect AI system decision can still create a safety risk.
Automotive-grade components do not replace application-specific safety assessment and system validation. Cybersecurity is also very important and we are incorporating it into our product strategy, along with other requirements like Europe’s Cyber Resilience Act. This is why real-time control and intelligence close to the point of action are important in robotics.
ELE Times: How have key robotic architectures evolved over the years?
Allan Lagasca: The first robots were centrally controlled machines. Today centralized processing and distributed intelligence work together. The central processor provides overall coordination for the system, while time-critical decisions, such as adjusting grip, can be made locally. This is important because even a few milliseconds of delay can affect the precision of coordinated joint movements. Humanoid robotics is still emerging, so we want to stay close to customers at this stage.
ELE Times: Humanoid robots are fast becoming socially visible. How do you see this technological progress augmenting human life across the globe? What are the future trends in humanoids we can look forward to?
Allan Lagasca: Humanoid robots can augment human capabilities by taking on repetitive and dangerous tasks in manufacturing and logistics, with potential applications in healthcare and daily life. Looking ahead, space exploration could extend this role to supporting astronauts, inspecting equipment and preparing infrastructure in hazardous environments.
Progress will depend on greater autonomy, dexterity, energy efficiency and reliable operation. More integrated modules and lower costs could accelerate adoption, but battery life, safety, regulation and public confidence remain critical. The goal is not simply more capable robots, but trusted systems that compliment people on Earth and beyond.
ELE Times: Tell us more about high-precision surgical robots and the complexity involved in their application.
Allan Lagasca: There is virtually no room for error in surgery. A surgical robot must replicate the surgeon’s handling and move the tools with a “natural” and smooth motion desirable for passing through the surgical site, while preserving vital information at all times even during constant motion. That requires accurate current sensing to maintain torque and to detect any faults as early as possible. ST’s robotics portfolio provides building blocks that can support these functions, spanning from the intelligent motor driver to current sensor, motion sensor and microcontroller for real-time control.
ELE Times: Describe the penetration of robotic applications in the medical industry. How do you address the criticality required in this domain?
Allan Lagasca: Medical robotics is a growing area within Smart Industrials. ST also addresses established rehabilitation and assistive applications, including prosthetics, exoskeletons and assistive robots. Our collaboration with NVIDIA supports physical AI development across humanoid, industrial, service and healthcare robotics. For systems interacting with patients and clinicians, safety is paramount.
Our building blocks support system-level safety and cybersecurity, while local processing can limit sensitive data transmission, depending on system design. We help medical device companies focus on innovation, while application-specific safety and medical-device validation remain the manufacturer’s responsibility.
ELE Times: With Artificial Intelligence making its way into every sphere of application, what are the innovations we can expect robotics to undergo in the foreseeable future?
Allan Lagasca: Robotics will see closer integration of AI, semiconductors, sensors, power electronics, software and mechanics. AI will require more diverse real-world data to support robotic learning and reliable operation. Edge intelligence will enable faster responses, while local processing—as in our camera demonstration—can help protect privacy.
Privacy, safety and cybersecurity must be critical design priorities throughout AI development and deployment, supported by responsible data handling and system validation. Integrating more functions into fewer components will also help reduce costs and improve efficiency, making capable, trusted robotics more accessible.
ELE Times: Describe the integration of ST’s technologies such as microcontrollers, sensors, and actuators into robotics systems.
Allan Lagasca: Sensors provide a robot’s “eyes” and “balance”; microcontrollers and processors act as its “brain”; motor control and power devices act as its “muscles”. ST’s portfolio spans more than 500 parts, covering approximately $600 of a robot’s bill of materials. Our module-to-system strategy integrates these technologies through reference designs and module partners, connecting hands, joints and cameras within a coherent architecture. We start with application requirements: for one customer’s 6DoF robotic hand, integrating six motors on a 6 cm × 6 cm board required a more compact system-level design using ST’s latest compact system-in-package motor drivers.
ELE Times: Finally, how does STMicroelectronics intend to shorten the path from concept to deployment for professional service robots?
Allan Lagasca: Our Robotics Segment Strategic Program is part of ST’s strategy to address the robotics market, advancing a module-to-system approach to help customers move from concept to deployment. Rather than simply selling individual components, we are bringing more of the building blocks together through reference designs and working with module partners. With integrated modules and demonstrated designs for hands, joints and cameras from ST and its ecosystem partners, customers can avoid building everything from scratch. This can simplify integration and manufacturing and help reduce system cost.
Our collaboration with NVIDIA aims to further shorten this journey by helping developers simulate and test robotics systems before building the physical hardware. The collaboration includes work to integrate ST sensors and motor-control solutions with the NVIDIA robotics ecosystem, including NVIDIA Isaac Sim, to support design and simulation before real-world deployment. In a market that is evolving this quickly, we want to be early and stay close to customers so we can help move promising concepts towards deployment faster.
The post Driving Motion Precision: How Physical AI is Converting Data into Action appeared first on ELE Times.