BI-DIRECTIONAL VISITOR SENSOR

Abstract

Knowing how many people are present inside a room, hall, shop, or building at any moment is useful for safety, energy management, crowd control, and business planning. Many public spaces have a maximum safe capacity, and exceeding it can create risk during emergencies. Lights, fans, and air conditioners are frequently left running in empty rooms, wasting electricity. In most places the number of people inside is either not known at all or is estimated by staff at the entrance. This paper presents a Bi-directional Visitor Sensor that counts people entering and leaving through a doorway automatically and uses the live occupancy count for monitoring and control. The proposed system uses a pair of infrared or time-of-flight sensors placed a short distance apart across the doorway and connected to an ESP32 microcontroller. When a person passes, the two sensors are interrupted in a particular order, and the controller uses this order to decide whether the person is entering or leaving. The count of people inside is increased for each entry and decreased for each exit, and the total number of entries and exits is also recorded for later analysis. The live occupancy count is shown on an LCD or seven-segment display at the entrance. When the room is empty, the controller switches off lights and fans through a relay module, and it switches them on again when the first person enters. When the occupancy reaches a set limit, an indicator at the door turns red and a buzzer sounds, informing staff and visitors that no more people should enter until others leave. The limit can be set separately for each room according to its size, seating, and fire safety rules. Visitor data is sent to a cloud platform using MQTT over Wi-Fi. A web dashboard shows the current occupancy of each monitored room, hourly and daily visitor trends, peak times, and alerts. Managers can set capacity limits remotely and receive mobile notifications when a space becomes overcrowded. The data can support staffing decisions in shops, scheduling in libraries and laboratories, and occupancy-based energy saving in offices and classrooms. Because only counts are transmitted, the amount of data sent is very small and the system works well even on a weak Wi-Fi connection. Testing at a classroom doorway showed that the sequence-based direction logic, combined with timing checks and debouncing, counted single-file movement accurately and handled common situations such as a person stopping in the doorway or turning back. The system is low in cost, does not use cameras, and therefore does not record any personal images. Future work includes handling of people walking side by side using sensor arrays or depth sensing, integration with building management systems, and linking multiple doors to track occupancy across an entire building.