Deep Robotics, a leading embodied AI and robotics company, unveiled its inaugural humanoid robot model the Dr.01 at the World Robot Conference that opened in Beijing, marking the startup’s efforts to diversify from quadrupedal robotics into the thriving humanoid robotics space.
During the five-day event, which runs through August 25, Hangzhou-based Deep Robotics also demonstrates a series of capabilities of its Embodied AI-powered quadrupedal robot.
Through AI and big data training, future humanoid robots are expected to assist humans in performing repetitive, high-precision tasks, safeguard personnel safety in hazardous environments, enhance human productivity in industrial settings, and improve the quality of life at home.
The Dr.01, developed by Deep Robotics, is an “AI plus” intelligent physical agent. Equipped with the firm’s proprietary lightweight J60 joints and high-power J100 joints, it boasts highly flexible movement capabilities, adapts to complex environments, and integrates sensing/perception abilities and powerful autonomous learning capabilities.
Deep Robotics will conduct a series of AI training sessions on the Dr.01 to advance the evolution and progress of “AI plus” intelligent agent technology.
Among all the types of bionic robots, quadruped robots are pioneers that stand out with more mature technologies and applications. They are already being used in power stations, factories, construction sites, underground tunnels, and utility corridors.
With the rapid development of AI, quadruped robots will gain more powerful learning, sensing/perception, decision-making, and planning abilities.
At this year’s World Robot Conference, Deep Robotics showcases the autonomous “dodging” skills of the X30, its flagship large-sized quadruped, for the first time.
The robot can recognize/detect suddenly approaching humans and obstacles, autonomously avoiding/navigating them to prevent collisions.
This makes interactions between the robot dog and humans safer, minimizes accidents during long-term operations, and improves the quality of long-term maintenance work.
Fostering the robot-human interaction and reducing the need for manual control has always been critical in robotic applications.
During the conference, Deep Robotics demonstrates the X30’s “lead-follow” function, allowing operators to control the machine’s direction and speed with a leash.
Under the follow mode, the robot can autonomously adapt to various random terrains, significantly enhancing the efficiency of human-robot collaboration in tasks such as outdoor transportation and material handling.
Deep Robotics’ Lite3 robot dog showcases advanced biomimetic jumping. By autonomously learning from animal movement patterns, it has no specific requirements for launch position, posture, terrain, or direction, significantly boosting its mobility efficiency in complex environments and enhancing its adaptability in natural settings.
Deep Robotics says it is a leader in Embodied AI and application innovation.
Its self-developed legged robots have been applied in various scenarios such as power station inspections, factory surveillance, corridor patrols, emergency rescue, fire detection, and future scientific research.
The robots have been deployed to the underground corridors for the duration of the 19th Asian Games in Hangzhou in 2023, participated in a Singapore National Grid project, and took part in earthquake relief and explosive detection exercises.
Deep Robotics currently engages in long-term deep collaborations with industry giants such as State Grid, China Southern Power Grid, Baosteel, and Fluke.
