Popular Keywords
News
[Insight] Analysis of Humanoid Robot Vision Systems and Opportunities for Taiwanese Companies
According to TrendForce, with AI empowering hardware devices, the applications for robots have expanded significantly. Beyond manufacturing, humanoid robots can be applied in general consumers’ work and life environments. The multifunctional humanoid robot, suitable for both manufacturing and service sectors, is attracting significant market interest. Companies such as Tesla, NVIDIA, and OpenAI are actively investing in the development of related products and services, recognizing the significant commercial potential in this sector.
Humanoid robots are capable of operating and learning to perform a wide range of tasks within environments designed for humans. This capability eliminates the need for manufacturers to develop multiple single-purpose robots or to modify existing environments and workflows, making humanoid robots highly attractive to the market.
In June 2024, Tesla announced plans to deploy humanoid robots in its electric vehicle factories, targeting the deployment of 1,000 robots by 2025. At its October 2024 product showcase, Tesla further demonstrated humanoid robots interacting with humans, showcasing their innovative potential in both manufacturing and service sectors.
From a statistical perspective, with major robot manufacturers gradually achieving mass production by 2025, TrendForce projects that the global humanoid robot market value could surpass USD 2 billion by 2027. The market is anticipated to grow at a compound annual growth rate (CAGR) of 154% from 2024 to 2027.
In addition to advancements powered by AI technology, improvements in sensor performance are also propelling the development of humanoid robots. The integration of AI with robotic vision, force feedback, and tactile hardware solutions enhances robots’ ability to adapt to diverse physical environments, respond to varying conditions, and engage in self-learning. As a result, sensors and related hardware devices are anticipated to experience significant market growth.
Mainstream Machine Vision Solutions
Among related techniques, machine vision solutions are especially crucial. In its latest report, TrendForce provides an in-depth analysis of the development of machine vision systems.
A machine vision system typically consists of five main modules: optical imaging, image sensors, image processing, input/output (I/O), and display/display unit. The optical imaging module uses lenses to project spatial information from the object side onto the image side. The image sensor module is responsible for converting optical signals into electrical signals. The image processing module, powered by a CPU system or dedicated chips, extracts information using complete processing solutions and algorithm libraries.
Machine vision provides substantial advantages over human vision, including superior capabilities in quantification, grayscale resolution, spatial resolution, and observation speed. By employing cameras, lenses, light sources, and lighting control systems, it collects detailed data on target objects. Machine vision systems integrate visual control mechanisms, intelligent image processing software, and algorithm libraries to analyze and process graphical information. Through the seamless collaboration of hardware and software, machine vision equips robots with advanced visual capabilities.
From the perspective of strategic deployment by leading manufacturers, Boston Dynamics and Tesla both introduced groundbreaking humanoid robot technologies in 2024. Boston Dynamics’ humanoid robot, Atlas, represents the pinnacle of complex robotic motion capabilities. Equipped with an advanced visual system, it delivers real-time stereoscopic environmental imagery, enabling autonomous navigation and obstacle avoidance. The company incorporates a range of mainstream vision sensing technologies into its humanoid robot.
In 2024, Tesla introduced the third-generation Optimus humanoid robot, marking significant technological advancements. Unlike other manufacturers that integrate various depth sensing technologies, Tesla leverages its expertise from autonomous driving and employs a fully in-house-developed FSD (Full Self-Driving) vision system. The new generation Optimus features Tesla-designed sensors and vision systems, achieving substantial improvements in weight and walking speed. The visual system builds upon Tesla’s expertise in autonomous driving, significantly enhancing the system’s real-time performance and improving its adaptability in complex environments.
NVIDIA’s promotion of its robotic learning platform is driving growth among Taiwanese vision system solutions suppliers, says TrendForce
The investments by leading companies not only inject significant momentum into the industry but also create new development opportunities for Taiwanese companies. Notably, the quality of data obtained through machine vision will be a key factor in determining the efficiency of robotic learning.
Leveraging their manufacturing strength and technical expertise, Taiwanese companies are actively positioning themselves in the visual solutions market. Many have partnered with NVIDIA to benefit from its platform’s market share during the early growth phase of the humanoid robotics industry.
Renowned for its expertise in AI vision software, Solomon has partnered with NVIDIA. The company offers over 300 applications, including deployments in Mercedes-Benz assembly lines and precision medical equipment manufacturing. These achievements showcase its strong capabilities in AI vision technology.
As a leading company in collaborative robots, Techman Robot combines AI and vision technologies to develop intelligent robots with built-in AI engines. Its no-code modular programming environment allows users to operate robots without any coding, significantly lowering the barriers to robot application development.
Yuan High-Tech Development has launched edge computing products based on NVIDIA Jetson platform. Its Pandora platform is designed to meet the control requirements of various types of robots, targeting use cases in retail and logistics. AVerMedia has collaborated with NVIDIA to advance applications in smart retail and smart cities, with its AI solutions expected to drive sustained revenue growth.
Canon is integrating optical technology with AI, aiming to launch robotic vision modules by 2025. These modules are primarily designed for applications in the medical and intelligent transportation sectors, targeting orders from Europe and the U.S. This marks a shift for Taiwanese manufacturers, transitioning from component suppliers to providers of comprehensive solutions.
Taiwanese companies, leveraging their strengths in key components such as servo motors and AI servers, have been actively developing solutions in high-value areas like machine vision and AI computing, showcasing their technical expertise and innovation. With the expansion of application scenarios for humanoid robots, Taiwanese manufacturers are expected to play an increasingly pivotal role in the global supply chain.
Read more
- [Insight] CES 2025 Spotlight: From NVIDIA Cosmos to Next-Gen Robotics – AI Innovations Showcase
- [News] Samsung Speeds up Humanoid Development to Offset Business Headwinds, Challenging Tech Giants
(Photo credit: NVIDIA)
