Intelligent air cooler using handy foot sensor person sensor

Intelligent air cooler using handy foot sensor person sensor

an introduction

Have you ever thought, how easy and simple it is to enable your devices with Edge AI without having to go through a cumbersome process of integrating a library or owning software engineering resources for ML Stacks? Yes, we have tools to make this process less difficult. However, what if the sensor could provide extracted features to the image data directly? Wouldn’t that be fun?

People Sensor Useful sensor is an artificial intelligence (edge) device/sensor that is able to detect a person in the camera frame and generate details about who (you can assign IDs to multiple people) and where the person is in the frame. It does all the heavy lifting of capturing, identifying a person, computing the NMS algorithm, and delivering indispensable data via I2C on demand through enough simple APIs.

With every other appliance getting smarter in this era, why not add smarts to our old summer friend’s air coolers? The Air Buddy from Crompton is specifically designed for use in the kitchen. It was supposed to provide airflow directed towards the cooker. This will not have any effect on the burner flame and therefore will not affect the cooking process. Although it has good airflow and swing features, it lacks proper cooling if it’s moving around the kitchen.

So, I thought about adding Vision to our Blind Air Buddy. Now, he can detect the presence of the person – “turn on the fan”. Track the person’s movement throughout the frame and adjust the blades accordingly to provide the user with efficient cooling. All these upgrades by just one sensor.

a job

The goal here is to upgrade the air cooler and adjust its blades to direct airflow towards the user in the tire. To add a vision to the current project the general procedure is as follows:

  • Take pictures from the vision sensor.
  • Process it on the MCU.
  • Implement edge-AI algorithms to first detect if a person is in the frame and secondly to calculate the relative coordinates of the person in the frame.
  • Use this generated data to position the servo motors.

However, it’s easier than said, especially for the first three steps. Person Sensor does the heavy lifting for the first three steps and provides the user with APIs to easily integrate the unit into the solution. Internally, the person sensor assembles an image sensor and a small microcontroller into a single board on which the user can manipulate any other sensor, giving it power and retrieving information through a standard I2C hardware interface. The sensor also supports creating an interrupt on the TP1 pin when a person is detected. This could be an external trigger to wake the MCU from sleep mode and start processing the data.

The air cooler has been modified by adding two servomotors for vertical and horizontal movement. Due to the mechanical structure of the coolant, we need to limit the range of motion of the servo motors so as to avoid any mechanical stress and breakdown of the motors.

Using the visual information from the person’s sensor, the centroid of the bounding box is calculated. These centroid coordinates are then mapped to the actual position of the servo motors. Also depending on the confidence rating, we control the relay that turns the air cooler on or off.

Hardware build:

Seeed Studio Xiao Ble is used to collect all data from the sensor and control the servo motors accordingly. Using expansion boards exposes all the pin needed in the orchard form factor and also keeps the system neat and tidy. Since the person sensor supports the I2C protocol and has a qwiic connector, using a grove to qwiic connector, it plugs into one of the I2C grove connectors on the expansion board. Xiao’s power supply is capable of running two servomotors. Hence, both motors are connected to Grove connectors with pins A0 and A7. You can change it as you wish. The relay, which is responsible for turning the air cooler on and off, is connected to pin D10.

Thin malleable steel bars were used to connect the vertical panels to the servo. For horizontal movement, a circular disk with a crank pin was connected to the servo motor. The motors were held in position using some nuts, bolts and a glue gun. After the mechanical design was fixed, the entire system was connected using large jumper wires to keep the system clean and tidy. Part of the air cooler’s power supply cable has been cut in half to enter the relay unit to control the power supply.

For demonstration purposes, the system is powered from a USB power bank connected to the Xiao MCU. However, power can also be drawn from the air cooler as well. Here are some shots of the designed devices:

Software build:

Since most of the calculations are done by a person’s sensor, the main code is quite straightforward. The program collects metadata of the person’s position in the frame through I2C and calculates the centroid from the same information. Then the relative coordinates are set to the actual engine position to direct the airflow. Confidence and person detection status are also displayed on the monochrome screen of the expansion board that shares the I2C bus with the person sensor.

The mbed I2C request from() is not working properly as expected. Thus, we will need to add the following code segment in a file wire File in the Arduino library:

size_t arduino::MbedI2C::requestFrom(uint8_t address, size_t len, bool stopBit) {
char buf[256];
rxBuffer.clear(); //Add This.

The software is built with PlatformIO, however, it should work fine on the Arduino IDE as well. You can follow This guide To set up the Xiao BLE environment in the Arduino IDE. Here are the links to the project repository and dependent libraries:

proof

Here is the moment of truth. Here is a playlist of videos showing the smart air cooler powered by person sensor:

System walkthrough:

Internal work:

Cooler Action Part One and Two:

future scope:

It’s a promising device and has a lot to offer that I haven’t explored yet. For example, the sensor on the device can be trained to detect many specific people with different IDs and will provide the bounding box information for each person detected.

The sensor opens up a plethora of opportunities to bake intelligence into existing devices.

For more information about the sensor, you can contact Pete Warden

#Intelligent #air #cooler #handy #foot #sensor #person #sensor

Leave a Comment

Your email address will not be published. Required fields are marked *