SHENZHEN, China, June 17, 2020 /PRNewswire/ -- The Global Top 5 wearable brand, Amazfit's mother company, Huami (NYSE: HMI) , made its second wearable AI chip Huangshan-2 public on its first Innovation Conference "AI to Decode Future". Besides, multiple hardware and a series of Big Data and AI solutions regarding healthcare were announced. Ramesh Jain, the industry famed the Father of Multimedia computing was announced as chief technical advisor of Huami AI Research Institute.
The Wearable AI Chip, Huangshan-2(MHS002) Launched.
As the platinum member of RISC-V, Huami has wide-ranging technology accumulation and investment layout in the field of semiconductors. In 2018, Huami launched the world's first RISC-V open-source instruction set wearable processor, Huangshan No.1 (MHS001), featuring four core artificial intelligence engines -- cardiac biometrics engine, ECG, ECG Pro, and Hearth Rhythm Abnormality Monitoring Engine.
After 450 days, Huami unveils its next generation of AI chip, the Huami Huangshan-2(MHS002). Based on the RISC-V architecture, Huami Huangshan-2(MHS002) features high computing efficiency and low power consumption. Huangshan-2 detects atrial fibrillation 7 times faster than Huangshan-1 and 26 times faster than similar algorithms. Huangshan-2 also has on board the Always On (AON) sensor mode with ultra-low power consumption thanks to the NPU (Neural-network Processing Unit) and the C2 co-processor. Theoretically, it can reduce the overall power consumption of Huangshan 2 by 50%, making a longer battery life for products.
Huangshan-2(MHS002) is scheduled for mass production in Q4 of 2020. New wearable devices equipped with the Huangshan-2 will be hopefully available on Huami's products in the first half of 2021.
BioTracker™ 2, the Second-generation PPG Bio-tracking Optical Sensor by Huami
Huami has been developing its own PPG bio-tracking optical sensor in the past years, which are known by its high accuracy. Compared with its predecessor, the newly launched, BioTracker™ 2 supports five biological data engines (RealBeats™, OxygenBeats™, SomnusCare™, ExerSense™, and huami-PAI™), making it the most versatile and precise biosensor ever developed by Huami.
Five Key Data Engines Contribute to Huami's AI Health Management Platform
To let human pursue the full health rights as the most, Huami's Systematic AI Health Management Platform incorporate the AI chipset, Huangshan-2(MHS002), the new PPG sensor BioTracker™ 2 and also five key index and data engines.
- RealBeats™ 2, the advanced AI bio data engine for heart rate monitoring. It eliminates the noise interference to heart rate signals during workout and is capable of monitoring atrial fibrillation for 1.87 times as long as the last generation at night and 6.64 times in the day. it achieved automatic AI-based detection of AVRT and frequent PVC through a heart health Big Data model.
- OxygenBeats™, Huami self-developed oxygen data AI engine. The success rate of is 100% in contrast to less than 90% seen in similar products. Compared with the results of professional oxygen analyzers, the average error is only 1.67%, reflecting accuracy superior to that of most wearable wrist devices for blood oxygen detection. OxygenBeats™ has applied to the follow-up visits with recovered COVID-19 patients conducted by Nanshan Zhong's medical team. Smart watches equipped with OxygenBeatsTM are expected to be available in Q3 of this year (2020).
- SomnusCare™, an AI-driven biological data engine. The accuracy of sleep data detection exceeds 80%, and it is possible to detect naps over 25 minutes with nearly 100% accuracy. Furthermore, the OxygenBeatsTM, through sleep state analysis and blood oxygen saturation detection, can recognize the sleep apnea syndrome (SAS), an "invisible killer" to human health, and warn the user to take prompt measures.
- ExerSense™, an AI recognition engine for motion patterns based on sports big data. This engine can match motion models in real time by detecting data with the motion sensor and heart rate sensor on Huami's wearables, and then intelligently determine the exercise mode of a user. At present, ExerSense™ can automatically detect 19 exercise modes, including walking, running, cycling, and swimming, which cover 95% of daily exercise scenarios of users. Users do not need to perform tedious manual operations, but can enjoy senseless selection of smart exercise modes.
- huami-PAI™, an explicit PAI score that can reflect the user's heart rate data, together with the daily motion time and other health data, allowing users to monitor their workout and heart healthy status. huami-PAI™ can guide users to create highly personalized health assessment systems using private bio data such as age, gender, and resting heart rate. huami-PAI™ is based on the HUNT Fitness Study, which helps to reduce the cardiovascular mortality rate and improve life expectancy. The study, led by Professor Ulrik Wisloff at Faculty of Medicine of the Norwegian University of Science and Technology, took 35 years with more than 230,000 participants.
Ramesh Jain appointed the Chief Technical Advisor of Huami AI Research Institute
Mr. Wang Huang, the Company's Chairman and CEO, announced the establishment of Huami AI Research Institute. Ramesh Jain, a renowned AI expert, Father of Multimedia Computing, Professor at University of California, Irvine, and Founder of UCI Institute for Future Health was appointed as the chief technical advisor of the Huami AI Research Institute.
In the first half of 2020, Huami has set up three joint laboratories: The Smart Wrist Wearable Device Joint Lab with the team of Zhong Nanshan; the Track and Field Joint Lab with the Chinese Athletics Association; and the Brain-Computer Intelligence Joint Lab with the Institute of Advanced Technology, University of Science and Technology of China.
 Huami ranked the top 5 in both global watch shipment and market share, according to data from the International Data Corporation (IDC) Worldwide Quarterly Wearable Device Tracker
 This is not a medical device and is not intended for use in the diagnosis or monitoring of any medical condition