Parents will sleep more peacefully thanks to Angee. They will be advised in case the baby is sleeping in dangerous positions through an alarm clock. Moreover if the baby starts crying the system will produce sweet sounds or project lights in order to calm the baby. In fact these functions can be also activated by parents during the day if they has to do little activities in the house and cannot keep an eye on him for few minutes. In short the system will alarm the parents providing relevant information (according to the context),detect baby's position and soothe crying baby’s, contributing to an overall safer and calmer environment.

Purpose and Scope

Angee it's a system that exploits a set of sensors, which are able to detect if the baby is in dangerous position and if the baby is crying, placed on the cradle of the baby and a set of actuators that tries to calm him/her down during day and night. The calming part is composed of projection and music. If the system is not able to tranquillize him/her or he/she is in the wrong position the user is advised through a notification on his phone. The user can disable some calming functionalities.

AmI steps

AmI step Description
Sensing sound sensor to check the baby's cry, cameras
Reasoning recognize baby position, if he/she is crying/in danger
Acting calm the baby with projector and sounds; alarm the mother if the baby is in danger
Interacting the mother can enable/disable the projector/sounds
AmI feature Description
Sensitive able to sense baby's problem (see above)
Responsive Able to respond to the needs of the baby and to act on the environment
Adaptive the system is able to understand if the baby is crying => calm the baby, if he is in danger=> alert the mother
Transparent The system behaves in an "embeded" way. The user does not have to know the details of the implementation to use it.
Ubiquitous sensors are in the room and collaborate to analyze the problem
Intelligent (see open issues)


USER: the tutor of the baby (parents in general)
BABY: the newborn in the cradle


Not visible: we assume that the baby’s face is covered (for example: baby on tummy with the face against the sheets) or baby outside cradle.
user: the tutor of the baby (parents in general)
projector: a simple system of lights and shadows

Functional Requirements

Functional area Description Priority
1 the system is able to monitor baby’s position through face recognition by using 2 cameras
FR.1.1 Face: The system continuously checks if the baby's face is visible.1
FR.1.2 Cradle-check:The system continuously checks if the baby is inside or outside the cradle.4
FR.1.3 Not_visible_alarm: If the baby's face is not visible and the baby is inside the cradle, it sets the "FACE NOT VISIBLE" alarm: sends a notification to the user.1
FR.1.4 Not_visible_alarm: The "FACE NOT VISIBLE" alarm is not send more than once every 30 seconds.1
2 the system is able to sense the baby’s cry and it will react if the baby is crying
FR.2.1Cry_check:The system continuously checks if the baby is crying1
FR.2.2Cry:The system assumes that the baby is crying if the sound is louder than the ambient noise level in db1
FR.2.3Cry_Reaction:If the baby cries for more than 5 seconds, the system turns on calming music in the speakers and the wireless LED lighting in red.1
FR.2.4Project:The lighting is used to project star shapes in the wall.3
FR.2.5Stop:If the baby is not crying for 30 seconds, the music and lights will be turned off.2
FR.2.6Limit:If the baby keeps crying for more than 10 minutes (in the presentation this will be set to 1 min,for us to be able to show it), the system turns off the music and the LED.3
FR.2.7Switch_On/Off:The User must be able to switch on/off the music or LED, at any moment, via physic buttons.2

Non Functional Requirements

Non Functional area title description area
NFR1 Compatbility All the code of the project is open source. Usability
NFR2 Language The system will communicate in English Usability
NFR3 Distance The sound machine is at least a foot (30 centimeters) away from the child's head. Safety
NFR4 Sounds The sound machine should produce low-pitched sounds. Safety
NFR5 Electricity the system must be plugged in Efficiency
NFR6 Supported system the system supports smartphones Usability
NFR7 Where the system can be used in every room of the house as the product can be transported Portability
NFR8 Color The LED color projected at night should not be blue Safety

Open Issues

forsee how the baby will react to the system
make the system understand that some sounds/videos calm the baby more than others in order to reuse baby's favourite more often
implement a mechanical system of rotating toys to entertain the baby
do an application in order to enable/disable calming functionalities


Cry detection: A baby's own cry is 10 times louder than a hair dryer.
What light colors will calm down a baby?
Don’t use blue light at night.Some flurescent lights are sold with coatings that can shift the color away from blue wavelengths.
Use dim red lights for night lights. Red light has the least power to shift circadian rhythm and suppress melatonin.
What musics will calm down a baby?
The sound machine can not go higher than 75 Db. Around 65 to 75.
The sound machine at least a foot (30 centimeters) away from the child's head.
Dr. Harvey Karp, Use low pitched sounds
Low pitch rumbly sounds are better for a baby's sleep but for promoting sleep, he says, aim for softer and lower-pitched.

Hardware architecture

Icon All the communications and decisions go through a Raspberry pi device, our one and only computational node.


Cameras are located over the cradle, in order to see if the baby’s face is visible, the video streaming is sent to Raspberry.
Microphone is located near the baby’s head, in order to stream the sounds to Raspberry that will calculate if the baby is crying.
Speaker is located near the cradle and it’s controlled by Raspberry.
Lights are located next to the cradle, projecting shapes to the ceiling in colours controlled by Raspberry.

User interface and devices :

Instapush on the mobile phone where the user, connected with the local server, recives notifications.

Hardware components

Sound audiocontroler, HEXIN
Multimedia Microphone, ewent
Phillips Hue, Phillips
Raspberry Pi 3 Model B, Raspberry
Raspberry camera module, Raspberry
Sound speakers, Goal Zero
1 breadboard, 2 330 ohm resistors, 2 led diodes, 2 switch, connectors

Software architecture

Software Purpose Location
Instapush send alarms to the user and is connected to the Raspberry Smarthphone
Facial Recognition elaborate the data from the cameras Raspberry
DB calculator calculate the power in db of the signal received from the microphone Raspberry
LED controller set specific colors acording to the data retrieved from the microphone and cameras Raspberry
General Controller gathers all the data and is responsible to send alarms via the local server to the mobile application Raspberry

Software components

All the following libraries are implemented using Python:
Instapush version 1.0.7
Control the LED using Philips Hue API / Rest Communication with the hue Bridge via Phue library.
Facial recognition/Image manipulation via OpenCV 2.4.13 library.
Decibels calculator via Analyse 1.0.0 and Pyaudio 0.2.8 libraries.
We also use: RPi.GPIO 0.5.11 to control the pins;
Pycurl and StringIO 7.5 to interract with Instapush;
Requests 2.7.0 to control the light;
Numpy 1.11.0, Thread 4.5.0 and time 15.3 to create the code.