Position: Home > News > capital
Using emg signals to identify gestures, aoyi wants technology iteration and content development to g
Core hints :The concept of AR/VR is not new, but the relevant market development is still slow. According to ni hualiang, CEO of aoy

Release date:2019-11-13

Browse number:226

 The concept of AR/VR is not new, but the relevant market development is still slow. According to ni hualiang, CEO of aoyi technology, in addition to the lack of content, the existing human-computer interaction experience is also a major obstacle to its development. Currently, the mainstream AR/VR interaction still relies on gamepad to realize, and users need to use gamepad or touch screen next to the head display for control, which contradicts the immersive feeling emphasized by AR/VR.
Aoyi technology sees the inherent advantages of market pain points and emg signals, and launches a gesture interaction scheme based on emg signals. Ni told 36kton that when the human body wants to do something, the nerve stimulation generated by the brain reaches the muscle fibers in the form of an electric current, which is medically known as a muscle electrical signal. Therefore, the emg of the arm is captured and collected in a circle, and then the gesture instructions can be derived reversely with the self-developed deep learning algorithm.
Ni hualiang introduced that the gesture control scheme based on emg has obvious advantages over the traditional computer vision gesture technology scheme in several aspects. Is A low cost, based on computer vision solutions need to deal with large amounts of data, chip costs mainly include depth image sensor module, can perform complex computing unit such as the Cortex of A data processing and pattern recognition. A, determines the overall cost is relatively high, its meaning in addition to independent research and development of science and technology of electrical sensors, data processing and computing unit USES is the low cost of conventional architecture (M chip, at present the overall product costs around $18 to 20 in the future cost of large quantity of postpartum will control within 15 dollars. Second, low power consumption, myoelectricity is direct drive reasons for action, and electromyographic signal is low frequency signal, effective spectrum within 2 KHZ, after acquisition of multiple channels electromyography together also can't more than 20 KHZ, with less interference information, thus ensures that reduce the computational complexity, from the methodology system of the overall power consumption can be effectively controlled within 20 ma. Third, the degree of freedom is high, not affected by the environment (light, shielding, image perspective, etc.).
In July, aoi introduced the GFORCE armband, an electromyographic gesture controller, which has built-in 8-channel medical-grade electromyographic sensors and a 9-axis motion sensor. When worn on the arm, sensors close to the skin pick up emg signals under the skin, and machine learning algorithms match them to the action. Since the product was launched not long ago, only a small number of participants' data have been collected. At present, the accuracy of gesture recognition reaches about 95%. Later, with the growth of users and training samples, the recognition accuracy will increase accordingly.
Using emg signals to identify gestures, aoyi wants technology iteration and content development to go hand in hand
There are 6 predefined gestures preset in GFORCE, and users can also modify these 6 gestures, collect customized gestures, and then train and learn the machine. Ni hualiang told 36kr that the number of gestures that can be processed is determined by both cost and user experience. At present, the positioning of the product is of low cost and low power consumption. Therefore, the processing power of the chip is limited. If the number of gestures is increased, a relatively large delay will be generated, which is not conducive to user experience. At the same time, too many or too complex gestures are difficult for users to remember and use. Six gestures are enough to meet the needs of ordinary users. If the cost is increased, it can also realize the capture of very subtle hand movements and multi-gesture recognition, and its application scenarios may be more in the industry field.
The next generation product of aoyi technology will further improve the user's wearing experience. The wearing part will be changed from the arm to the wrist, while the size will be smaller, which is similar to a bracelet or a watch. Ni said the wrist is close to the tendon and the emg signal is weak, so the accuracy of the sensor and algorithm is required.
The benchmarking company ni introduced to aoyi was Thalmic Labs, a Canadian startup whose product is MYO gesture control armband. Thalmic Labs just announced a $120 million series B round of funding in September from Intel capital, amazon Alexa fund, Fidelity Investments Canada, and others. Ni hualiang told 36kton that they have carried out six iterations of GFORCE's emg sensor, and have done a lot of tests on the device structure and layout as well as the selection of materials for the sensor to touch the skin. Therefore, compared with the limitation that Thalmic Labs MYO product requires users to warm up and use it in 2-5 minutes after wearing, the sensor of GFORCE has better anti-noise performance and strong consistency of signal processing during use, so it can be used immediately after wearing. The price of GFORCE is about 600 yuan, which is more competitive than 36kr's previously reported thousand-yuan DTing wristband.
Aoyi technology will work with b-end users and d-end developers in the next six months to a year to help develop more interesting content and applications by providing core algorithms and sensors, open SDKS and apis. NiHuaLiang think AR/VR is a very clear product applications, but the application of the electrical control technology is not limited to the AR/VR field, it can replace the existing control methods used in the intelligent TV, can also be applied in medical treatment, to do mechanical prosthetic control, as well as the application in the field of sports, in professional training on muscle balance, fatigue and use efficiency of monitoring.
VR/AR has established a deep partnership with ruiyue information, leciang technology, aqua technology and other companies. Our partners in healthcare include hanlin technology, heishang technology and Enabling The Future 3D printed prosthesis website. At present, GFORCE can support gesture control of more than 20 games on the VR integrated machine of ruiyue and lesang, or connect PC to make gesture PPT page turning, browse the website, and interact with fruit ninja and other motion sensing games.
Ni hualiang, CEO of aoyi technology, graduated from electronic engineering department of fudan university, with 16 years of industry experience and 7 years of overseas work experience. Before starting his business, he worked in qualcomm Canada. Yu zhou, CTO, has 17 years of experience in software development and management in the consumer electronics industry.
Aoyi technology is seeking pre-a round of financing, which will mainly be used to expand its r&d team and iterative products.

Friendship link    | International Artificial intelligence

News Exhibition Video Product Enterprise Picture Technology
Liaoning ICP No. 234234234 -1 | Liaoning No. 1102342342030919 public security