A method for estimating bus congestion based on mobile phone sensors

Abstract

The invention relates to a method for estimating the degree of bus congestion based on mobile phone sensors, that is, using the inertial sensor that comes with the mobile phone, using the idea of ​​group intelligence to collect data, and then performing information on the movements and postures of passengers on the bus under the framework of hierarchical learning identification, so as to achieve accurate bus congestion estimation process. The present invention can use the inertial sensor attached to the mobile phone to collect sensor data from passengers getting on the bus until they reach a static state, use these data to identify the passenger's action posture information, and estimate the congestion degree of the bus through these information.

Description

technical field [0001] The present invention designs sensor applications, crowd perception, machine learning, action recognition, gait recognition, passive perception, layered learning, bus congestion estimation, etc. in computer science, especially a bus congestion sensor based on mobile phone degree estimation method. Background technique [0002] In recent years, as the sensor equipment equipped on mobile phones has become higher and higher, coupled with the continuous in-depth development of the field of machine learning, gesture recognition relies on fewer and fewer sensors, and at the same time, the accuracy of prediction is getting higher and higher. Relying on gesture recognition, we can achieve indoor positioning, daily sports data statistics, and can combine location information or communication signals to track a person's behavior. Predicting the degree of congestion of the bus is a relatively new application in the direction of action recognition. If the degree...

AI Technical Summary

Problems solved by technology

[0003] To obtain the degree of congestion in an area, it can be estimated by obtaining the total number of people in this area. The existing methods of counting people are all based on images or videos, which means that these methods rely on camera equipment, although these methods are in image Or under the premise that the video is clear, there is already enough accuracy, but the low-cost camera equipment cannot guarantee the clarity of the obtained image data in the case of poor light, and the high-cost camera equipment is obviously impossible

So people-based methods are not practical

The situation of people on the bus can also be counted by the data returned by the pressure sensor. This method will not be affected by environmental factors such as light and humidity, and the result is very accurate. However, if one of the sensors is broken, the number of people in the entire bus will The statistics will gradually tend to be too much or too little, and in the end there will be completely wrong results

Images