Jietong Huasheng Lingyun makes the A U robot "can listen and speak"
On April 8th, Jietong Huasheng was invited to attend the “Smart Child Road†service robot conference jointly organized by Cixing and Ayou Culture. The general manager of the company, Wu Weidong, represented the company’s pistachio robot jointly established by Cixing and Ayou Culture. Signed a strategic cooperation agreement.
The U-Yuzhi Children's Robot (hereinafter referred to as Rabbit Intelligence Robot) released this time is a family service robot that can accompany children to learn and play. Jietong Huasheng provided the Lingyun ring microphone array and Lingyun intelligent voice technology for the rabbit intelligence robot, so that the rabbit intelligence robot can not only accurately identify the owner's words, but also can tell the cute children's voice and let the children and rabbit intelligence robots voice. The interaction is smoother and more natural.
As a highlight of the product, voice interaction is the main way for children to communicate with rabbit intelligence robots. Due to the family environment where the rabbit intelligence robot is located, there are often environmental noises such as household appliances and indoor entertainment. The child may also be close to the rabbit intelligence robot, so the noise recognition capability and far field recognition ability of the speech recognition technology are higher. Claim. In this regard, Jietong Huasheng provides Lingyun ring microphone array for rabbit intelligence robots, which can directly capture the children's voice, and accurately identify the words of distant children through series of techniques such as noise reduction, speech enhancement and echo cancellation. At the same time, the adoption of Lingyun speech synthesis technology allows the rabbit intelligence robot to not only say cute children's voices, but also to sing clear and bright and pleasant songs with children's voices, so that the rabbit intelligence robot can "speak and sing" and greatly enhance The interactive experience of children and rabbit intelligence robots.
Lingyun ring microphone array allows robot to accurately identify owner's discourse
In order to enable rabbit intelligence robots to accurately obtain voices from all directions within the family, Jietong Huasheng has specially launched the Lingyun ring microphone array. Unlike the car microphone array, because the car device only needs to pick up the occupant's utterance within 180° of the device, a linear microphone array can meet the demand, while the family range is more open, the speaker will move, and the sound source may be in various directions. A circular microphone array is needed to locate and acquire voice information in the 360° range.
Lingyun ring microphone array can accurately calculate the angle and distance between the sound source and the array module according to the time difference of sound arrival, realize the positioning of the target sound source, and use the spatial domain filtering characteristics of the microphone array to form in the target speaker direction. The pickup beam picks up only the beam's signal, suppressing noise and reverberation (reflected sound) outside the beam.
Since the sound energy is attenuated as the propagation distance becomes longer, the array module enhances the overall picked speech beam energy by organically superimposing the sound signals picked up by the plurality of microphones. At the same time, the echo cancellation technology is used to filter out the sound emitted by the machine's own audio equipment, and to input a clearer and pureer sound for the speech recognition engine.
Due to the inevitable pick-up of some environmental noise during the pickup process of the microphone array module, Jietong Huasheng models the environmental noise and trains with a large amount of real noise data to optimize the speech recognition algorithm. Lingyun speech recognition technology with strong anti-noise ability can easily cope with common environmental noise. At the same time, the deep neural network (DNN) algorithm used in this technology can use the accumulated speech data training and optimization algorithm in the user's use process, and continuously improve the anti-noise ability, so that the robot can accurately identify the owner's discourse.
Lingyun speech synthesis allows the robot to "speak and sing"
In addition to using the Lingyun ring microphone array and speech recognition technology, the rabbit intelligence robot also uses the Lingyun Huasheng's Lingyun speech synthesis technology, which can express cute, tender and natural words, which makes people feel like a six or seven-year-old. The child speaks. As the earliest enterprise engaged in the research and application of speech synthesis technology in China, Jietong Huasheng has launched the Lingyun speech synthesis technology representing the industry's top level as early as 2001. After more than ten years of technology accumulation, and in thousands of companies The application practice and running-in, Jietong Huasheng constantly optimizes the text analysis model and perfects the rhythm planning, so that the synthesis effect of Lingyun speech synthesis technology is true and full, and the MOS value has reached 4.6. The intelligibility, naturalness and fluency reach the industry. Top level.
Lingyun speech synthesis technology has fully supported Chinese, Mandarin, English, Cantonese, Uighur, Japanese, Korean, Russian, German, French and other 21 languages, covering more than 90% of the world's countries and regions, Sino-British mixed reading, Yueying The mixed reading is natural and fluent, fully adapting to the current multilingual speech synthesis requirements, allowing the robot to “speak†a smooth, natural discourse.
Recently, Jietong Huasheng has launched Lingyun singing synthesis technology by relying on its deep accumulation in the field of speech synthesis for decades. According to the song information, the technology allows the robot to sing sweet, cute, clear and translucent, and sings like a real person. It successfully realizes the leap-forward progress of the machine from "speaking to talking" to "singing and singing", so that the robot can say Will sing." Letting the robot learn to sing, not only enhances the fun of the robot service, but also means that the robot has taken a big step in the personification and service experience.
Lingyun platform is fully open to the robot industry
As the leading enterprise in China's artificial intelligence industry, Jietong Huasheng launched the first all-round artificial intelligence open platform at home and abroad, Hyuncloud.com. The Lingyun platform is fully open to the industry and provides the industry with speech synthesis, speech recognition, OCR, handwriting recognition, face recognition, voiceprint recognition, fingerprint recognition, semantic understanding, machine translation, etc. in a “cloud+end†manner. Azimuth artificial intelligence capabilities.
Since the beginning of 2015, Jietong Huasheng has announced that Lingyun platform is fully open to the Chinese robot industry, and has successively been hundreds of hundreds of pistachios, Zhongzhi IOT, Xinsong, Zhaowei, Cobos, Avatar, Chino, NAO, etc. Domestic and foreign robot manufacturers provide a full range of Lingyun artificial intelligence capabilities, and fully support the Chinese robot industry to create a real robot that can listen, speak, think, and judge. A large number of robots using Lingyun capabilities have been implemented in batches, and more large-scale tests have been carried out in financial, telecommunications, energy governments, business offices, shopping malls, etc., which have received extensive attention from industry and users. With praise.
In recent years, along with the explosive growth of the artificial intelligence industry, Jietong Huasheng relies on the deep knowledge of artificial intelligence technology and accurate prediction of the development of the robot industry. With the concept of cooperation and win-win, it does not compete with robot manufacturers, nor does it compete with robot manufacturers. The sense of cooperation in the market has won wide support from domestic robot manufacturers. At present, it has established a solid partnership with most domestic major robot manufacturers. Lingyun artificial intelligence has become unique in the Chinese robot industry, almost everywhere!
In the future development, Jietong Huasheng will continue to rely on the research strength of Tsinghua University's artificial intelligence research, comprehensively respond to and support the national innovation and development strategy, comprehensively promote the comprehensive docking of the Lingyun platform and the Chinese robot industry, and cooperate with all Chinese robot manufacturers. Together to create the glory of the Chinese robot industry.
Linear Encoder,Digital Linear Encoder,Draw Wire Sensor,1500Mm Linear Encoder
Jilin Lander Intelligent Technology Co., Ltd , https://www.jilinlandertech.com