The Most Reasonable Elevator Scheduling Strategy Based on AI

The judgment of the full staff inside the elevator only stays at the judgment of whether the bearing quality is overloaded, ignoring the evaluation index of the congestion degree. When the people (including goods) carried in the elevator are crowded but not full, the outside personnel will do the ladder operation, and the elevator will stop when passing the ladder floor. Because the elevator is already crowded, many escalators choose not to enter. Wait for the next elevator. If such a scene occurs during the peak hours of the boarding and there are many floors, the elevator will stay on each floor one by one, thereby lengthening the waiting time of the passengers inside and outside the elevator, reducing the operating efficiency of the elevator and increasing the energy consumption of the elevator.

In order to solve the above problem, we introduce a congestion recognition module in the elevator car. The module consists of a camera, a processor, a congestion recognition algorithm, and so on. After the image information collected in the camera is processed by the recognition algorithm, the current congestion degree parameter of the car is obtained, and it is determined whether the parameter reaches the threshold as another condition for the elevator to be full.

The core part of the congestion identification module is the congestion recognition algorithm. The instant video in the car obtained by the camera is processed by the congestion recognition algorithm to process the current video and output the current congestion value.