A device for counting the passenger flow is introduced

To the problem of making plans, arranging schedules and forecasting for transit agency, the passenger of public traffic is of importance. Compared with the existing application technology, a new passenger statistics solution is introduced. A kind of self-designed contact sensor pedal is used to sample footprint data and identify the direction when passenger gets on and off. The principles of kinesiology and cluster analysis are proposed in this method to complete pattern classification. It is of high accuracy, and low cost. The test results show that the accuracy rate can be above 95%.