Intelligent Yuan Interactive Service Robot

Personification configuration, human factors design

Referring to ergonomic design, the height to size ratio is close to that of humans, with a technological appearance and elegant yet powerful wheel lines

Adhering to the principles of ergonomics, HRl (Human Robotic Interaction) design is more closely aligned with the natural human interaction experience

Universal human skeleton configuration and modular design, supporting personalized appearance expansion capabilities, meeting the diverse needs of user scenarios.

Multi mode interaction, intelligent experience

Based on speech modeling, hyper personification synthesis technology, full duplex dialogue capability, and end-to-end model deployment, the real-time dialogue of robots is made more lively and approachable.

Building a customized business knowledge base for enterprises based on the language big model+RAG (Retrieval Augmented Generation), supporting precise marketing language, business consulting and guidance, and exhibition explanation, making information transmission more efficient and intuitive.

Based on a multimodal human model, the multi-channel information processing capability makes human-computer interaction easier and more comfortable; By integrating sound source localization, facial and lip detection algorithms, the accuracy rate of sound pickup in noisy and high noise scenes can reach 96%, and the facial wake-up rate can reach 99%.

Based on the Action GPT model, hundreds of simulated natural human actions can be dynamically generated based on human voice commands and intentions, and light task execution can be performed.

Autonomous mobility and reliable walking

Based on the HlMUS (High performance Multimodal Mapping System) 3D-SLAM algorithm Vectorflux control algorithm, with low latency and low drift performance, it supports L4 level autonomous movement in complex and dynamic environments.

The mass-produced self-developed integrated joint with high reliability, lightweight, and ultra-high slot full rate winding technology has a higher peak torque of 430N. m, a motor torque density of 50Nm/kg, and dynamic performance. The joint is reliable and supports long-term stable walking.

Combined with RTMOF (Real Time Robust Motion Framework) nonlinear model predictive control and reinforcement learning algorithms, as well as an efficient dynamic programming engine, it can quickly generate globally optimal trajectories and continuously optimize motion strategies, with strong anti-interference ability, achieving agile, reliable, and flexible walking in uncertain environments.

Multiple safeguards, secure and worry free

360 ° LiDAR+6 high-definition cameras fusion perception, achieving panoramic without blind spots; Combining open set semantic information to support synchronous construction of dynamic and static semantic environments and targets, achieving intelligent obstacle avoidance.

Three layers of security monitoring are implemented at the business layer, system layer, and hardware layer. The computing center and security center are mutually redundant, with dual control from the security center to the actuator, achieving PLd level security protection.

Convenient operation and minimal maintenance

Support remote control, mobile phone, and computer multi platform control

The multifunctional standby station supports transfer, standby, and charging, reducing the frequency of personnel intervention.

Supports two types of energy replenishment methods: battery quick change and charging, to meet the needs of various scenarios.