Dear ALL Clients,

Due to the National Holiday, our office will be temporarily closed from Oct 1st to Oct 7th, 2024. We will resume work on Oct 8th, 2024.
For urgent issues, feel free to contact  +8613656000291 (Tank) or +8617700087816 (Jasmine).

Best Regards,
Xiamen Kegan IoT Technology Co., Ltd.

​Ⅰ Requirements
Taking an elevator in a high-rise building in a big city, the following situations can happen:
(1) Someone on certain floors presses the elevator, but then suddenly leaves for something else, causing the elevator to stay, which is very common in life.
(2) It is also common in life that the elevator is already packed with people but still stays on the floor where someone pressed the elevator but can no longer carry anyone else.
(3) Some high-rise buildings will exist 4 to 6 elevator devices, multiple button operation, will lead to the existence of multiple elevators to the same floor, but at this time the passengers have been carrying the first to arrive at the elevator to leave, after arriving at the elevator belongs to an empty trip.
(4) Some people will press the wrong elevator but will not cancel it in time, but the elevator will still lift to the wrong floor after stopping at the right floor, which is more common when going up the stairs.
Although the current elevator is smart enough, it has not completely solved the above problems, as well as other problems that the blogger has not thought of.
 
Ⅱ Solution
It just occurred to me that millimeter-wave radar for human detection has been studied for more than 10 years, especially in the last 5 years the outbreak of a large number of human presence detection radar, multi-target tracking and head counting radar, etc., their costs have been very low. If the characteristics of millimeter wave radar can be used, it should be able to solve the problem of elevator intelligent linkage to a certain extent:
(1) If the elevator is already stuffed with people, it will only go down but not up;
(2) If there is no one outside the elevator door, then don't stay;
(3) If there is no one inside the elevator, then keep silent.
This not only saves electricity, but also accelerates the passengers' riding experience, so that they will not be annoyed because the elevator is always doing some useless work, especially during the peak period of commuting.
 
Ⅲ Implementation Program
(1) Install a headcount radar inside the elevator, which requires high-resolution radar. Do better with cameras.
(2) Each floor to install a low-cost human presence detection radar can be, 24GHz can be, after all, a lot of floors, the deployment of the cost is very high.
(3) Finally, link the data from the radar with the elevator control system, while ensuring the stability of the system and the correct functionality of the system.

The application of millimeter-wave radar sensors in self-driving mainly includes real-time monitoring of the surrounding environment of the vehicle.

​By emitting millimeter waves and receiving reflected signals, information such as the distance, speed, and angle of target objects can be detected. In autonomous driving, millimeter-wave radar can help vehicles achieve functions such as adaptive cruise control, automatic emergency braking, and lane departure warning, improving driving safety and comfort.

Indoor millimeter wave radar sensors seem to be small in size and simple in hardware structure, but it is still very difficult to do in practice, and the complexity of the radar should not be trivialized because of its small size. Specifically manifested in:
(1) The algorithmic scheme has the requirements of the arithmetic power to be low but also requires the algorithmic performance to be very good.
(2) The hardware cost should be low, need to cooperate with the low arithmetic processor to complete the balance between real-time and accuracy.
(3) Engineering landing needs to face a lot of potential problems that need to be solved, such as a variety of interference, blind zones, false alarms, omissions, and other issues.
(4) The indoor environment is complex, and the targets have different characteristics, which requires good performance of radar processing algorithms.
(5) The consistency of the hardware is also important. Millimeter wave radar sensor is very small, but some manufacturers make a board, the consistency of performance is not very good, some boards have better signal quality, some boards have poorer signal quality. In addition, power spurious is also a problem.
(6) Human presence detection is simply the human presence detection function. For example, human presence detection, while combining gesture recognition, head count, sleep monitoring, fall monitoring, and so on. To unify and integrate these functions is actually a relatively difficult thing.
(7) Some indicators and precision data are difficult to achieve.

​Millimeter-wave radar sensor applications have gradually expanded and are now mainly focused on automotive radar , transportation security, indoor homes, and other popular areas. Demand is also growing in other niche areas. Although both are millimeter wave radars, the processing method and focus differ in different ways. For example, in automotive radar sensors :
 
(1) Most of the targets are cars belonging to the rigid body target, do not need to pay too much attention to the details of the target's movement information, such as wheel rotation;
 
(2) Generally, an automobile radar needs to detect targets within 300m at the farthest, and needs to realize stable tracking of targets and estimate the direction of target movement;
 
(3) Even in the face of pedestrians and other non-rigid targets, automotive radar does not need to pay attention to the target's specific attitude changes, such as pedestrian detection does not need to pay attention to the specific posture of the person (running, walking, falling), as long as the car is told someone and their movement direction.
 
However, indoor radar and automobile radar sensor are significantly different. Indoor radar sensor ​has many scenarios, such as human presence sensing, respiratory heartbeat detection, gesture recognition, gesture recognition, fall detection, number of people, etc., as well as based on these basic functions of the expansion of the application, such as smart lights, air conditioning, TV, toilet seat and so on. In these scenarios, it is necessary to sense information about the human body or human limbs, and academics have categorized the different levels of motion on the human body into three types, as shown in Figure 1.

Figure 1: Classification of Human Body Motion