Team Members

Theory: In my college, we have a lot of electricity poles here and there, plus there are no major admin methods to work with them rather than to just on or off them all at once. Also, it is been speculated that by 2050 almost 70% of the world population will be living in cities which would then comprise of locality level colonies. So we have to make great use of our resources and this project will be one amongst many.

Overall Hardware Required: A raspberryPi, rasberryPiCamera, GPS tracker, astronomical time, RFID reader, power supply, 4G Connection, PCB boards, and other miscellaneous items.

Idea: My idea is to make use of the existing technology to better the overall locality level experience of an individual using smart light poles which would then have an impact on the whole society in general.

  1. USING ELECTRICITY EFFECTIVELY: A major part of electricity is being wasted on using LEDs at their full current bias but at night with no one in the locality, there is no reason to make them work in their full potential. Using the Google Weather API and using a Gps along with raspberryPi we can make a get request to the google weather API to give the data in JSON format. The data given out has the following classes:

    Weather Condition Duty Cycle Output

    CONDITION_CLEAR 20%

    CONDITION_CLOUDY 50%

    CONDITION_FOGGY 100%

    CONDITION_HAZY 100%

    CONDITION_RAINY 70%

    CONDITION_STORMY 80%

    SUNRISE + OffsetValue 0%

    CONDITION_UNKNOWN *To be predicted from past data

    Now we can change the duty cycle of the PWM signal that is fed into the Led to make it work thus changing its brightness with current Weather Conditions.

  2. BRIGHTNESS BASED ON OBJECT DETECTION: A lot of localities doesn’t have anybody on roads but we have a lot of power consumption as the Led Lights work on their full power. I propose a system that will evaluate and do object detection using opencv3 and if any object/vehicle is seen whose position changes with time, then it makes a post request to an endpoint. This endpoint has an algorithm that calculates the speed of the vehicle. Then sends a response to the raspberryPi installed. The data sent will be: Category{Speed; Timestamp;}

    The category will contain Car, Bikes etc. This raspberryPi will then send this data to the next raspberryPi installed and brightness of the pole is decided.

    Example: Suppose the two poles are 3 m apart and the speed of the vehicle is 50km/hr. Then the time taken by the vehicle to parse this distance is 0.21 s. So the brightness will change in (0.21 s – error_offset) of each pole from a nominal 20% to 100% and back to 20%.

  3. CLEANLINESS OF LOCALITY: In India cleanliness is also a major problem and most of these workers don’t work how they are expected to work. So what I propose is to use a raspberry Pi with piCamera to send the pictures of the roads to a server that is running sends a request to the locality manager to look at these pictures and make a tick mark if he/she is convinced with the work and a cross if he/she isn’t through a reactNative app. Using the data collected over a month we can pay these workers effectively and for the work that they have done.

    VALUE BINARY EQUIVALENT TICK 1 CROSS 0

  4. HELP BUTTON: A major cause of problems in the world is theft/bad-deeds and also lack of immediate police action. I want to make an RFID based system that has a HELP button over it. Whenever a resident clicks on ‘this’ button, a post request will be sent over to the server which then will inform the security guards via SMS present at the location that an Emergency situation is there on this latitude and longitude. The location coordinates will be sent by the client-app installed as soon as the RFID tag is authenticated by the RFID reader. In this way, we can make sure that an emergency situation is been taken care of at the earliest.
  5. LIVE SECURED CAMERA READ: Using raspberry Pi and piCamera we can make a secure CCTV system but at a very low cost. The camera will send real-time camera_frames to a remote server via web sockets implemented with either flask or Django channels. We can also manipulate the camera_frames to apply object detection of vehicles only using a pre-learned CNN that will save all the vehicles that passed through a particular location and also their timestamp in the video. Using this data, we can see whether a particular vehicle passed through a particular location or not. And this can be used to quickly catch culprits rather searching through a whole day long CCTV footage.

Inspiration

Whenever I come back from the library I wonder why are the lights always in full brightness mode and this made me think of this idea. Plus when ElectroBoom mentioned about this contest I researched and documented all the stuff. Plus I think practical knowledge is what I should aspire for so I want to make this model.

Voting

Voting is closed!