Video detection and analysis cannot guarantee the privacy and security of users; smart phones cannot meet various needs due to different sizes, built-in sensors, and 
processing performance; and sensor detection methods currently mainly use 
accelerometer sensors and 
human motion-based  Different algorithms of acceleration characteristics are used to detect the fall state. However, some daily high-intensity sports, such as jogging and sitting down quickly, will also produce a large acceleration feature similar to falling. Using the acceleration sensor alone for 
fall detection, the collected data is relatively simple.  It is not enough to fully reflect the changes in 
human body posture, and it is easy to cause many false alarms
 On the other hand, fall monitoring is based on post-fall treatment and cannot achieve the effect of prior prevention
[0005] Moreover, some monitoring devices on the market are bulky, not only unable to provide convenience for the elderly who need to go out, but also unable to monitor the physical condition of the elderly in real time, nor can they monitor the real-time location of the elderly in real time, while other portable devices have single functions,  It is not enough to detect diverse activities. In addition, if a single remote 
server goes down, the entire 
system will be paralyzed. Therefore, there is a great need for a portable 
sports injury monitoring system to protect the health of the elderly in real time
[0006] According to the above analysis, the current sports injury 
monitoring system mainly has the following problems: 1. The detection method is single, which will send 
false alarm information to the system, and the recognition accuracy rate is low
2. The alarm is issued after the sports injury, which cannot reduce the sports injury and lacks pre-judgment
2. The clustering using the 
attractor propagation 
algorithm is affected by the bias parameter, which is easy to cause misjudgment
Since the switching of the detection device 1 and the mode requires manual execution, the degree of intelligence is low
2. Judging by the acceleration threshold is easy to cause misjudgment
Due to the detection device 1, the judgment method of passing the threshold value, the 
false positive rate is high
2. The 
Bluetooth communication method is adopted, the communication distance is short, and the guardian cannot be notified in time
Since the division of the detection device 1 and samples will affect the subsequent detection, the data of squatting may be the same as the combined acceleration and 
angular velocity of falling
2. Use the calculation method of 
conditional probability to find the maximum value of the probability, and the 
noise has a great influence on the calculation
Because the detection device 1 and the 
big data platform require a large number of data samples as training, the amount of data in the initial stage will affect the judgment result
2. Using the similarity method is likely to cause misjudgment. If you sit down quickly, the system will also consider it as a fall state
Since the detection system 1, the 
processing flow of each data packet is trained through the model of the 
cloud server, the time for falling judgment is increased, resulting in the failure of timely rescue
2. Access to the network through WIFI, the scope of application is small
Because the monitoring system 1. The device is defined with a 
network interface module, it is not suitable for carrying out, and data cannot be transmitted as long as it leaves the 
network interface. 2. Read the 
ECG signal, and perform QRS 
wave detection, 
heart rate calculation and analysis of five kinds of arrhythmia.  Is a common diagnosis, can not be related to the analysis of the state of motion
2. The way of falling is judged by the acceleration threshold, which is easy to cause false alarms
Because the monitoring system 1. uses the device to send data to the 
mobile phone, it is easy to carry, but it cannot work independently without the 
mobile phone, and the portability is average.
2. The collected data are 
heart rate and 
blood pressure, which cannot reflect all situations and may easily cause false alarms
[0019] To sum up, the main deficiencies in the open patents and technical solutions related to the remote monitoring system include: 1. The function is single, or the acceleration threshold is used to judge, or the sole pressure method is used to judge, or a single ECG data is used.  Judgment, it is easy to cause misjudgment, and the recognition accuracy rate is low
2. The recognition speed is slow, and an 
alarm signal is sent after a fall is detected and a sports injury is received, and it is impossible to reduce or avoid sports injuries by predicting in advance
3. Lack of an effective and 
highly sensitive monitoring mechanism. After detecting damage, a single 
alarm message is sent to the guardian, which may easily lead to untimely rescue
4. Insufficient portability, unable to adapt to various occasions. For example, it can only be used indoors, or requires a 
network interface to use, or requires a 
mobile phone as a transfer
6. Lack of 
perception of the surrounding environment, single consideration of the feedback of the 
human body itself, lack of prediction of the environment