In 2018, we started the industrialization process with the formed technology and set up a company to promote it, which has been recognized by experts and peers in the industry. Now we hope to get the recognition from parents.
We compared the smart baby care products sold in the market, and there are mainly 4 kinds of plans for infant health monitoring:
- Intelligent wearable devices based on PPG or ECG. This product has mature software and hardware ecology. Represented by Apple Watch, it realizes all-round monitoring of exercise heart rate and blood oxygen. However, its working mode also limits the scope of use for infants and the elderly. Infants' skin is not suitable for wearing contact sensors for a long time, and the battery power life problems lead to such products cannot high-density collection and analysis of health indicators, and often cannot be put into use immediately because they forget to charge.
- Smart pillow or mattress device based on BCG piezoelectric sensor. It has the characteristics of non-touch close detection of mobile signals, which is suitable for bed care of patients. However, such sensors are highly affected by vibration and body movement, and are only suitable for breathing and heart rate monitoring during full sleep, rather than indicators for infants and the elderly when awake.
- The scheme based on PCR sensing, that is, microwave radar sensing. It has good penetrability at close range and can penetrate clothing and coverings such as quilts to monitor breathing and heart movements. But the product of this scheme on subjects such as chest strict focus detection position location, need to infants and young children or the elderly with the positioning of the bed when the direction of the radar wave can get good results, in addition many newborn parents of radar wave direct illuminate an infant may cause other problems concerned, such devices using fewer radar chip suppliers, Intellectual property and high manufacturing costs have limited the number and scale of applications of this microwave detection technology for infant and child health care products.
- Health care scheme based on CV machine vision. Based on large-scale machine vision technology and the popularity of cloud computing, these products upload raw video data to the Internet and hand over complex ARTIFICIAL intelligence calculations to cloud computing servers for analysis and feedback. Such technologies have some low cost advantage, users to buy the equipment price is low, but need to bear the cost of cloud computing brings long-term and additional Internet access fees, once disconnected with the Internet can not achieve its function of health monitoring, a lot of families are worried about the video upload data privacy security problem brought by the Internet.
When our potential customers put forward requirements to us for these problems, we made the following requirements for the functional positioning of this product based on these problems:
- Non-contact sensing scheme to avoid the use of children's skin allergy and discomfort;
- Use mature software and hardware solutions to reduce the difficulty and price of mass production;
- Adopt off-line edge computing scheme to realize no Internet connection requirement and privacy protection;
- Automatically detect the position and direction of the human body to be detected, reducing the necessary requirements for users to cooperate with the use;
- It can be linked with many household IOT devices to achieve multi-device query and alarm ability;
Children can communicate with their guardians by remotely viewing live videos and making voice calls，or sometime dad or mom is far awar from home,so you can coax your baby to sleep.
We thought it’s a good idea, so we quickly produced the first concept art.
Some functions of this product may have the same purpose as other products, but this is not the point. We consider that contact wearable devices may bring discomfort to babies, and some products need to use other accessories to achieve the function of intelligent image recognition. We break the rules, users in the process of using our products, do not need to use other auxiliary accessories. This product only needs a device +APP to complete the collection of physiological indicators of the detection object. Through intelligent image recognition, data such as heart rate, blood pressure, respiration and sleep posture can be obtained by non-contact detection. Through algorithm recognition, data within the normal range can be compared, thus prompting and warning can be given to users.
After a large number of intelligent face generation and video training software algorithm, the recognition accuracy reached 99.7%.
As a new team, although we have taken the road that other teams in the same industry have taken, we have also learned the advantages and disadvantages of rival products. We invested more budget in the software algorithm, and in the hardware, we had to adopt the common hardware module of the industry under budget pressure, and through algorithm transplantation and optimization, match the hardware level, so as to achieve the balance of software and hardware, and try our best to achieve the product functions originally planned by us.
Fortunately, after countless tests, our product has finally entered the stage of trial production after confirming the stability of the product function. At present, the industrial design drawing of the first version of the product is about to be completed, and our samples are coming soon!
This resulted in a product design solution: Ai baby monitor.