OPPO's new patent revealed: Using artificial intelligence to calculate the cause of a phone fall?

I often see people asking on Weibo, what is the one thing they must bring with them when they go out? Almost everyone answers that it is their cell phone.

Looking back at the evolution of mobile phone design in recent years, we can find that the appearance of mobile phones has changed a lot. Mobile phones have become more and more beautiful and good-looking, especially mobile phone screens have become larger and larger, and the screen-to-body ratio has become higher and higher, evolving from the pursuit of narrow bezels to today's full screen.

However, current mobile phones have a fatal problem, that is, the screen is too easy to break, and a careless drop may cause damage to the screen.

When a mobile phone falls and the screen is damaged, the repair cost is becoming increasingly high. This is a problem that everyone from manufacturers to consumers has to face.

Therefore, from a design perspective, it is an important science to rationally layout the structure of the mobile phone so that it can provide maximum protection for the screen. Correspondingly, the drop data research that provides data support for this cannot be ignored.

It is reported that OPPO has applied for a related patent technology, which uses artificial intelligence to use relevant data when the mobile phone falls to infer the cause of its fall.

as the picture shows:

For the falling process of the mobile phone, the relevant data may include the degree of grip of the user's mobile phone when it falls, specifically the amount of pressure on the mobile phone casing, and the direction and magnitude of the acceleration of the mobile phone during the falling process.

Furthermore, the two sets of data are recorded continuously - the continuous here means that the time interval is relatively small, which is actually still intermittent, not continuously recorded - to obtain the pressure and acceleration information of the mobile phone over a period of time. Then, when the mobile phone determines that it is in a falling state, by reading the data record at the last moment before the fall, the information recording the state of the mobile phone when the fall occurred is found, and the cause of the fall can be inferred based on this.

For the specific implementation method of inferring the cause of the fall, a trained artificial intelligence model is used. The so-called training means that a series of the above-mentioned fall data and fall causes are first input into the artificial intelligence model, and the model is allowed to perform forward calculations to obtain the relevant weights of the fall data to the fall causes, and then reverse calculations are performed to infer the fall causes based on the weights and fall data, and the weights are iteratively modified based on the results. After a certain iterative process, fairly reliable weight information is obtained. At this point, the model has been trained and can be used to infer the cause of the fall based on the fall data.

However, the causes of the fall need to be defined in advance, which may include: intentional fall, unintentional fall, negligent fall, etc.

After such an invention including a device and a method is implemented, the analysis of the cause of the fall can be realized. By summarizing this information of many users, the proportion of each of the above three causes of fall in the actual situation can be analyzed.

I personally think that if the proportion of accidental falls is relatively large in reality, then when designing the product, it is necessary to consider improving the grip and surface roughness of the phone to improve the user's grip on the phone in daily use, so that the phone is not easy to fall off the hand under normal circumstances. This can also largely avoid the occurrence of phone falls. After all, if a phone is always easy to fall off and the screen is broken, then I think no one is willing to buy this phone!