Mr Viral
Las Vegas, NV – January 3, 2025 – RoboSense, a pioneering AI-driven robotics technology company, hosted the world’s first AI Robotics Global Online Launch Event of 2025, “Hello, Robot.” This event marked a pivotal moment for the company, showcasing its strategic vision as a global robotics technology platform and unveiling groundbreaking achievements. Among the highlights were cutting-edge products that will debut at CES 2025, from January 7–10 in Las Vegas.
During the event, Mark Qiu, CEO of RoboSense, delivered a keynote, sharing the company’s journey and vision:
“This isn’t just a launch—it’s the start of a new chapter for robotics. Over the next decade or even longer, one day, embodied intelligent robots will coexist with humans, forming a new society of human-machine symbiosis.”
Building on a decade of technological innovation, RoboSense has demonstrated unparalleled expertise in LiDAR development and intelligent robotics. The launch event introduced RoboSense’s humanoid robotic prototype, a universal development platform that reflects the company’s in-depth understanding of intelligent robotics. This comprehensive platform lays the foundation for advancements in vision, tactile sensing, mobility, and manipulation solutions.
During the keynote, Mark Qiu highlighted the significance of LiDAR in robotics:
“LiDAR has fully entered the digital era! Much like the evolution from film cameras to digital cameras, this is a new chapter for LiDAR and robotics.”
He also emphasized the importance of incremental components in the future of robotics:
“Developing modules from scratch poses significant challenges in time, expertise, and cost. By providing advanced, reliable incremental components, RoboSense enables developers to accelerate innovation and adoption.”
Align with its goal to become the “global leader in robotics technology platforms,” RoboSense unveiled itsuniversal development platform for robotic components to accelerate the development of robotics components and solutions. This platform integrates three core areas:
Vision Solutions: RoboSense introduced the Active Camera, an intelligent ecosystem comprising sensor hardware, computational cores, and AI algorithms. By merging LiDAR signals and camera data, the Active Camera provides active environmental perception, overcoming challenges like lighting variability and visual illusions. Developers can use its SDK for seamless integration of mapping, localization, and obstacle avoidance, reducing development time and system complexity.
Manipulation Solutions: RoboSense showcased its second-generation dexterous hand Papert 2.0 with 20 degrees of freedom, incorporating hand-eye coordination to create a closed-loop system. This innovation significantly enhances versatility and complexity across a wide range of tasks.
Mobility Solutions: Building on its extensive automotive expertise, RoboSense has introduced an intelligent solution – Robo FSD to expanded into mobile robotics, enabling robots to achieve autonomous point-to-point mobility across various scenarios.
Q&A with Mark Qiu
Pandaily: How do you see the development of “embodied intelligence” over the past few years? Why has this field gained momentum recently?
Mark Qiu: The rise of embodied intelligence is driven by the gradual maturity of AI technologies. We call it “embodied intelligence” because it builds on AI foundations, enabling robots to actively perceive and respond to the external world. For instance, in some ways, a car can be seen as an embodied intelligent agent. Advances in automotive technology have made many critical components, like LiDAR, more mature, allowing us to focus on niche areas within robotics, including humanoid robots.
Pandaily: Specifically, what technological breakthroughs in embodied intelligence have occurred in the past two years?
Mark Qiu: We can divide embodied intelligence into two parts: the “upper body” and the “lower body.” The lower body focuses on mobility, such as point-to-point navigation in intelligent vehicles, especially high-speed mobility with robust external perception systems. This lays a solid foundation for embodied intelligence. The upper body emphasizes operational capabilities, such as applications in healthcare. Recent breakthroughs include better integration of upper and lower body solutions and the widespread adoption of incremental components. For example, LiDAR, which was scarce a decade ago, is now readily available. In our ecosystem, we focus on providing sensors and other incremental components to empower the industry.
Pandaily: From a market perspective, embodied intelligence doesn’t seem to have reached full maturity, especially in terms of commercialization. For instance, Elon Musk’s vision of a “robot girlfriend” still feels far away. What’s your take?
Mark Qiu: That’s true. If we set our sights on humanoid robots that exceed human capabilities, it will indeed take a long time. However, focusing on industrial, commercial, or even household robots, we already see many technologies being implemented. By integrating upper and lower body solutions, existing robots can solve highly specific problems. As incremental components mature, these robots have the potential to become much more intelligent. Like The products we just released has made revolutionary breakthroughs in the performance and accuracy of sensors.
Pandaily: From a technical perspective, what are the current challenges in achieving embodied intelligence?
Mark Qiu: In terms of mobility, while technologies in intelligent vehicles are quite advanced, challenges remain for rugged terrains, stair climbing, and seamless indoor-outdoor transitions. Enhancing durability and reliability is critical. For operational capabilities, creating a closed-loop system for hand-eye coordination, such as precise manipulation with dexterous hands, remains a significant challenge. Moreover, sensor technologies—including vision, touch, force perception, and electrical feedback—still need improvements in performance and precision.
Pandaily: Intelligent driving and embodied intelligence share a lot of overlapping technologies. How do you see the synergies between the two?
Mark Qiu: There are indeed many commonalities. For example, technologies developed for intelligent driving, such as point-to-point navigation and environmental perception, can be applied to robotics. End-to-end AI models, in particular, are making robotics more flexible by reducing the limitations of rule-based systems.
Pandaily: You mentioned that data is core to AI development. What are the specific challenges related to data in embodied intelligence? How is the industry addressing them?
Mark Qiu: Data is the biggest bottleneck. Unlike intelligent driving, embodied intelligence lacks sufficient operational data. The industry is currently addressing this through three approaches:
1. Real-world data collection: This is straightforward but costly and limited in scope.
2. Synthetic data: This involves combining existing datasets to create new scenarios.
3. Generated data: Using physical laws and simulations to generate virtual datasets.
A hybrid approach using synthetic data for pre-training and real-world data for fine-tuning is currently mainstream in the industry. This method balances cost and effectiveness.
Pandaily: At this year’s CES, which embodied intelligence technologies do you think are closest to commercialization?
Mark Qiu: Industrial and commercial robots are the most promising areas. Examples include intelligent forklifts, port automation, cleaning robots, and inspection robots, which are already seeing widespread adoption. In the consumer market, robots for tasks like lawn mowing, snow clearing, and pool cleaning are also rapidly growing due to their well-defined use cases and scalability.
Pandaily: How does RoboSense position itself within the embodied intelligence ecosystem?
Mark Qiu: We aim to be a robotics technology platform company, empowering three types of users:
1. Developers in the embodied intelligence field.
2. Practitioners in traditional industries seeking smart upgrades.
3. Technology enthusiasts.
Whether you need incremental components or complete solutions, we strive to provide the tools and support that you need.
Pandaily: Finally, how do you envision the development of embodied intelligence over the next decade?
Mark Qiu: The next decade will be a period of rapid growth for embodied intelligence. Key breakthroughs will come from advancements in computing power, algorithms, and data. The increased generalization capability of large models, the evolution of sensors, and the development of electric actuators will all contribute to the maturity of robotics. Although challenges remain, I am very optimistic about the future.
SEE ALSO: RoboSense Debuts 940nm-M3 LiDAR at CES 2024, Outperforming 1550nm Laser for Extended Range
GIPHY App Key not set. Please check settings