The voice interaction system independently developed by OrionStar has the industry-leading full-link far-field voice technology and a rich voice content matrix, and supports small, non-secret payments based on voiceprints.
Today, the fully self-developed Orion Voice OS technology has been applied to the products of many partners such as Xiaomi Xiaoai, Huawei AI speaker, Midea Xiaomei AI speaker, Himalaya Xiaoya speaker, Cheetah Mobile Xiaobao AI speaker, etc., supporting China's smart speaker market share exceeds 30%, becoming a widely used AI voice system. According to statistics, the Orion Voice OS technology has more than 30 million voice command requests per day, with millions of hours of far-field voice data accumulation, and cumulative activation of more than 100 million devices.
Self-developed robot-specific ring 6-microphone array, 360-degree full-field coverage, 150 million wake-up practice tests, 5 meters effective far-field interaction, customized for the scenes.
The self-developed face recognition algorithm has won the Champion Microsoft Million Celebrity Recognition Contest. It can accurately perform face recognition and follow-up, gender and age analysis, gesture recognition, object recognition and positioning, and support access control, payment, meeting reservations, and inspect application scenes and mobile applications deployment. At the same time, vision technology has been applied to service robot products to realize functions such as active visual interaction and assisted navigation and obstacle avoidance.
The self-developed indoor navigation system has industry-leading low-cost multi-modal navigation technology, multi-sensor fusion, accurate positioning, active obstacle avoidance, autonomous route optimization, all-day battery life, and automatic recharge when the battery is low.
Six-axis robotic arm: integrates industry-leading robotic arm vision and motion control algorithms, provides an intelligent teaching system, remote background monitoring system and rich artificial intelligence API. Around it, you will have the ability to build a powerful robot, and OrionStar will provide one-stop service to help launch the product.
It is equipped with an industry-leading robotic nervous system, and based on hardware-accelerated deep learning algorithms, it achieves extremely fast facial and human recognition, voice semantic customization capabilities, far-field recognition, wake-free, multi-round interaction, voiceprint technology and other industry-leading technologies. In addition, OrionStar and Rockchip Microelectronics jointly released the industry's first full-chain AI voice chip-OS1000RK. OS1000RK uses a low-power and high-performance CPU core-64-bit 4-core ARMCortex-A35, integrated CODEC (8-channel ADC + 2-channel DAC), can support 8 microphone arrays at low cost, its hardware voice detection module (VAD ) It can achieve a very low standby function, and a dedicated instruction set can make the operation of the neural network more optimized.