Supported by the powerful Tegra TX1 SOC visual computer, the binocular visual computing system of MYNT SDENO ROBOT can perform a 3D reconstruction of the environment in real time, recognize the environment, locate and avoid obstacles based on visual computing. The proprietary binocular camera in MYNT® EYE uses high-definition frame synchronization and carries a global shutter to refresh at a frequency of 60 fps for binocular frame synchronization at the chip level. Its built-in six-axis IMU sensor effectively reduces errors so that the robots can perceive their own trajectories. The patent for core components (pending approval) is structured to solve problems in obstacle avoidance and depth computing for white walls and glass surfaces.
Local real-time calculations based on CUDA acceleration; face recognition for simultaneous recognition of multiple faces and persons; personal identification with an accuracy of up to 85%; detection of facial similarities; identity comparison between photos and persons to allow enterprise customers to set up their security whitelist and blacklist.
With a computing power of up to 1T Flops, 256 Maxwell desktop graphics computing cores, and 8X64-bit CPU cores, this is the best system for deep learning, robotic vision, computer graphics, and GPU computing.
Developed in-house using a deep learning engine, appropriate expressions can be naturally mapped to corresponding scenes and dialog contents. In addition, character settings will be gradually added to make it easier for customers to adapt the system to their specific application scenarios and own custom attributes.
The knowledge base system is developed in-house and makes use of deep learning in computers to perform autonomous optimization based on user feedback during actual operations. The iFlytek speech recognition engine has a recognition rate of up to 98% for Chinese language and supports many Chinese dialects like Cantonese.
The excellent motor control algorithm allows robots to accelerate, decelerate and travel smoothly. The independent suspension system provides stability while the robot moves, and the proprietary wheels can easily deal with complex road conditions and can be used on various indoor and outdoor grounds.
Update with the cloud knowledge base easily with no training required to complete the update of all information to the robots quickly.
MYNT SDENO ROBOT only needs to go round the venues once to perform automatic environment recognition and mapping. Alternatively, you can talk to MYNT SDENO ROBOT while walking to help it create location tags, which is easy and convenient.
Just update the cloud knowledge base to update product communications, layouts, and shops, so as to save high training costs.
MYNT SDENO ROBOT is competent in many reception jobs like showing the way, picking up mails and packages, as well as receiving and guiding guests.